2019-04-12 06:36:43 +02:00
// Copyright (c) 2014-2019, The Monero Project
2018-04-10 06:49:20 +02:00
// Copyright (c) 2018, The Loki Project
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// All rights reserved.
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// Redistribution and use in source and binary forms, with or without modification, are
// permitted provided that the following conditions are met:
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// 1. Redistributions of source code must retain the above copyright notice, this list of
// conditions and the following disclaimer.
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// 2. Redistributions in binary form must reproduce the above copyright notice, this list
// of conditions and the following disclaimer in the documentation and/or other
// materials provided with the distribution.
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// 3. Neither the name of the copyright holder nor the names of its contributors may be
// used to endorse or promote products derived from this software without specific
// prior written permission.
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
// MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
// THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
// STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
// THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
2015-12-14 05:54:39 +01:00
//
2014-07-23 15:03:52 +02:00
// Parts of this file are originally copyright (c) 2012-2013 The Cryptonote developers
2014-03-03 23:07:58 +01:00
2017-11-25 23:25:05 +01:00
# include <boost/algorithm/string.hpp>
2020-10-24 00:49:42 +02:00
# include "epee/string_tools.h"
2014-03-03 23:07:58 +01:00
# include <unordered_set>
2020-12-01 01:56:43 +01:00
# include <sstream>
2018-07-21 03:27:13 +02:00
# include <iomanip>
2022-02-10 19:25:55 +01:00
# include <oxenc/base32z.h>
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
2019-10-07 19:23:16 +02:00
extern " C " {
# include <sodium.h>
2020-01-19 23:44:18 +01:00
# ifdef ENABLE_SYSTEMD
# include <systemd / sd-daemon.h>
# endif
2019-10-07 19:23:16 +02:00
}
2020-02-03 03:03:32 +01:00
# include <sqlite3.h>
2014-03-03 23:07:58 +01:00
# include "cryptonote_core.h"
2020-12-17 06:52:16 +01:00
# include "uptime_proof.h"
2020-05-30 06:06:16 +02:00
# include "common/file.h"
# include "common/sha256sum.h"
2017-09-14 05:39:37 +02:00
# include "common/threadpool.h"
2017-10-28 17:06:43 +02:00
# include "common/command_line.h"
2020-07-27 05:50:20 +02:00
# include "common/hex.h"
2021-02-08 06:31:05 +01:00
# include "common/base58.h"
2020-10-24 00:49:42 +02:00
# include "epee/warnings.h"
2014-03-03 23:07:58 +01:00
# include "crypto/crypto.h"
# include "cryptonote_config.h"
2021-06-18 01:06:51 +02:00
# include "cryptonote_basic/hardfork.h"
2014-09-30 21:24:43 +02:00
# include <csignal>
2017-09-10 18:35:59 +02:00
# include "checkpoints/checkpoints.h"
2016-06-16 00:37:13 +02:00
# include "ringct/rctTypes.h"
2015-03-25 16:41:30 +01:00
# include "blockchain_db/blockchain_db.h"
2022-04-29 01:51:14 +02:00
# include "blockchain_db/sqlite/db_sqlite.h"
2017-01-14 14:29:08 +01:00
# include "ringct/rctSigs.h"
2018-09-29 22:18:08 +02:00
# include "common/notify.h"
2017-11-29 23:53:58 +01:00
# include "version.h"
2020-10-24 00:49:42 +02:00
# include "epee/memwipe.h"
2018-07-21 03:27:13 +02:00
# include "common/i18n.h"
2020-10-24 00:49:42 +02:00
# include "epee/net/local_ip.h"
2022-09-13 08:57:56 +02:00
# include "logging/oxen_logger.h"
# include <fmt/std.h>
# include <fmt/color.h>
# include <oxenmq/fmt.h>
Change logging to easylogging++
This replaces the epee and data_loggers logging systems with
a single one, and also adds filename:line and explicit severity
levels. Categories may be defined, and logging severity set
by category (or set of categories). epee style 0-4 log level
maps to a sensible severity configuration. Log files now also
rotate when reaching 100 MB.
To select which logs to output, use the MONERO_LOGS environment
variable, with a comma separated list of categories (globs are
supported), with their requested severity level after a colon.
If a log matches more than one such setting, the last one in
the configuration string applies. A few examples:
This one is (mostly) silent, only outputting fatal errors:
MONERO_LOGS=*:FATAL
This one is very verbose:
MONERO_LOGS=*:TRACE
This one is totally silent (logwise):
MONERO_LOGS=""
This one outputs all errors and warnings, except for the
"verify" category, which prints just fatal errors (the verify
category is used for logs about incoming transactions and
blocks, and it is expected that some/many will fail to verify,
hence we don't want the spam):
MONERO_LOGS=*:WARNING,verify:FATAL
Log levels are, in decreasing order of priority:
FATAL, ERROR, WARNING, INFO, DEBUG, TRACE
Subcategories may be added using prefixes and globs. This
example will output net.p2p logs at the TRACE level, but all
other net* logs only at INFO:
MONERO_LOGS=*:ERROR,net*:INFO,net.p2p:TRACE
Logs which are intended for the user (which Monero was using
a lot through epee, but really isn't a nice way to go things)
should use the "global" category. There are a few helper macros
for using this category, eg: MGINFO("this shows up by default")
or MGINFO_RED("this is red"), to try to keep a similar look
and feel for now.
Existing epee log macros still exist, and map to the new log
levels, but since they're used as a "user facing" UI element
as much as a logging system, they often don't map well to log
severities (ie, a log level 0 log may be an error, or may be
something we want the user to see, such as an important info).
In those cases, I tried to use the new macros. In other cases,
I left the existing macros in. When modifying logs, it is
probably best to switch to the new macros with explicit levels.
The --log-level options and set_log commands now also accept
category settings, in addition to the epee style log levels.
2017-01-01 17:34:23 +01:00
2014-03-03 23:07:58 +01:00
DISABLE_VS_WARNINGS ( 4355 )
2017-02-23 21:15:52 +01:00
# define BAD_SEMANTICS_TXES_MAX_SIZE 100
2019-04-04 02:15:57 +02:00
// basically at least how many bytes the block itself serializes to without the miner tx
# define BLOCK_SIZE_SANITY_LEEWAY 100
2014-03-03 23:07:58 +01:00
namespace cryptonote
{
2022-09-13 08:57:56 +02:00
static auto logcat = oxen : : log : : Cat ( " cn " ) ;
static auto omqlogcat = oxen : : log : : Cat ( " omq " ) ;
2017-10-28 17:06:43 +02:00
const command_line : : arg_descriptor < bool , false > arg_testnet_on = {
" testnet "
, " Run on testnet. The wallet must be launched with --testnet flag. "
, false
} ;
2020-08-11 23:41:54 +02:00
const command_line : : arg_descriptor < bool , false > arg_devnet_on = {
" devnet "
, " Run on devnet. The wallet must be launched with --devnet flag. "
2018-02-16 12:04:04 +01:00
, false
} ;
2018-06-14 21:11:49 +02:00
const command_line : : arg_descriptor < bool > arg_regtest_on = {
" regtest "
, " Run in a regression testing mode. "
, false
} ;
2019-11-01 16:16:37 +01:00
const command_line : : arg_descriptor < bool > arg_keep_fakechain = {
" keep-fakechain "
, " Don't delete any existing database when in fakechain mode. "
, false
} ;
2018-06-14 21:11:49 +02:00
const command_line : : arg_descriptor < difficulty_type > arg_fixed_difficulty = {
" fixed-difficulty "
, " Fixed difficulty used for testing. "
, 0
} ;
2019-09-17 03:20:15 +02:00
const command_line : : arg_descriptor < bool > arg_dev_allow_local = {
" dev-allow-local-ips "
, " Allow a local IPs for local and received service node public IP (for local testing only) "
, false
} ;
2018-02-16 12:04:04 +01:00
const command_line : : arg_descriptor < std : : string , false , true , 2 > arg_data_dir = {
2018-01-21 16:29:55 +01:00
" data-dir "
, " Specify data directory "
2020-10-22 19:55:33 +02:00
, tools : : get_default_data_dir ( ) . u8string ( )
2020-08-11 23:41:54 +02:00
, { { & arg_testnet_on , & arg_devnet_on } }
, [ ] ( std : : array < bool , 2 > testnet_devnet , bool defaulted , std : : string val ) - > std : : string {
if ( testnet_devnet [ 0 ] )
2020-10-22 19:55:33 +02:00
return ( fs : : u8path ( val ) / " testnet " ) . u8string ( ) ;
2020-08-11 23:41:54 +02:00
else if ( testnet_devnet [ 1 ] )
2020-10-22 19:55:33 +02:00
return ( fs : : u8path ( val ) / " devnet " ) . u8string ( ) ;
2018-01-22 02:49:51 +01:00
return val ;
}
2018-01-21 16:29:55 +01:00
} ;
2017-11-30 16:35:52 +01:00
const command_line : : arg_descriptor < bool > arg_offline = {
" offline "
, " Do not listen for peers, nor connect to any "
} ;
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
const command_line : : arg_descriptor < size_t > arg_block_download_max_size = {
" block-download-max-size "
, " Set maximum size of block download queue in bytes (0 for default) "
, 0
} ;
2017-10-28 17:06:43 +02:00
static const command_line : : arg_descriptor < bool > arg_test_drop_download = {
" test-drop-download "
, " For net tests: in download, discard ALL blocks instead checking/saving them (very fast) "
} ;
static const command_line : : arg_descriptor < uint64_t > arg_test_drop_download_height = {
" test-drop-download-height "
2018-03-01 12:36:19 +01:00
, " Like test-drop-download but discards only after around certain height "
2017-10-28 17:06:43 +02:00
, 0
} ;
static const command_line : : arg_descriptor < uint64_t > arg_fast_block_sync = {
" fast-block-sync "
, " Sync up most of the way by using embedded, known block hashes. "
, 1
} ;
static const command_line : : arg_descriptor < uint64_t > arg_prep_blocks_threads = {
" prep-blocks-threads "
, " Max number of threads to use when preparing block hashes in groups. "
, 4
} ;
static const command_line : : arg_descriptor < uint64_t > arg_show_time_stats = {
" show-time-stats "
, " Show time-stats when processing blocks/txs and disk synchronization. "
, 0
} ;
static const command_line : : arg_descriptor < size_t > arg_block_sync_size = {
" block-sync-size "
, " How many blocks to sync at once during chain synchronization (0 = adaptive). "
, 0
} ;
2018-11-02 23:27:25 +01:00
static const command_line : : arg_descriptor < bool > arg_pad_transactions = {
" pad-transactions "
, " Pad relayed transactions to help defend against traffic volume analysis "
, false
} ;
2018-07-18 23:24:53 +02:00
static const command_line : : arg_descriptor < size_t > arg_max_txpool_weight = {
" max-txpool-weight "
, " Set maximum txpool weight in bytes. "
2022-05-16 22:55:05 +02:00
, DEFAULT_MEMPOOL_MAX_WEIGHT
2018-01-29 22:20:24 +01:00
} ;
2018-07-12 04:07:34 +02:00
static const command_line : : arg_descriptor < bool > arg_service_node = {
" service-node "
2021-04-08 18:42:33 +02:00
, " Run as a service node, option 'service-node-public-ip' must be set "
2018-07-12 04:07:34 +02:00
} ;
2019-06-04 06:15:16 +02:00
static const command_line : : arg_descriptor < std : : string > arg_public_ip = {
2019-06-27 08:03:22 +02:00
" service-node-public-ip "
2019-06-04 06:15:16 +02:00
, " Public IP address on which this service node's services (such as the Loki "
" storage server) are accessible. This IP address will be advertised to the "
" network via the service node uptime proofs. Required if operating as a "
" service node. "
} ;
2019-10-15 19:45:58 +02:00
static const command_line : : arg_descriptor < uint16_t > arg_storage_server_port = {
2021-04-08 18:42:33 +02:00
" storage-server-port " , " Deprecated option, ignored. " , 0 } ;
2019-10-15 19:45:58 +02:00
static const command_line : : arg_descriptor < uint16_t , false , true , 2 > arg_quorumnet_port = {
" quorumnet-port "
, " The port on which this service node listen for direct connections from other "
" service nodes for quorum messages. The port must be publicly reachable "
" on the `--service-node-public-ip' address and binds to the p2p IP address. "
" Only applies when running as a service node. "
, config : : QNET_DEFAULT_PORT
2020-08-11 23:41:54 +02:00
, { { & cryptonote : : arg_testnet_on , & cryptonote : : arg_devnet_on } }
, [ ] ( std : : array < bool , 2 > testnet_devnet , bool defaulted , uint16_t val ) - > uint16_t {
return defaulted & & testnet_devnet [ 0 ] ? config : : testnet : : QNET_DEFAULT_PORT :
defaulted & & testnet_devnet [ 1 ] ? config : : devnet : : QNET_DEFAULT_PORT :
2019-10-15 19:45:58 +02:00
val ;
}
} ;
2021-04-08 18:38:51 +02:00
static const command_line : : arg_descriptor < bool > arg_omq_quorumnet_public {
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
" lmq-public-quorumnet " ,
" Allow the curve-enabled quorumnet address (for a Service Node) to be used for public RPC commands as if passed to --lmq-curve-public. "
" Note that even without this option the quorumnet port can be used for RPC commands by --lmq-admin and --lmq-user pubkeys. " ,
false } ;
2018-09-29 22:18:08 +02:00
static const command_line : : arg_descriptor < std : : string > arg_block_notify = {
" block-notify "
, " Run a program for each new block, '%s' will be replaced by the block hash "
, " "
} ;
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
static const command_line : : arg_descriptor < bool > arg_prune_blockchain = {
" prune-blockchain "
, " Prune blockchain "
, false
} ;
2019-01-09 15:29:28 +01:00
static const command_line : : arg_descriptor < std : : string > arg_reorg_notify = {
" reorg-notify "
, " Run a program for each reorg, '%s' will be replaced by the split height, "
" '%h' will be replaced by the new blockchain height, and '%n' will be "
" replaced by the number of new blocks in the new chain "
, " "
} ;
2019-05-08 13:11:23 +02:00
static const command_line : : arg_descriptor < bool > arg_keep_alt_blocks = {
" keep-alt-blocks "
, " Keep alternative blocks on restart "
, false
} ;
2014-03-03 23:07:58 +01:00
2019-07-17 08:01:54 +02:00
static const command_line : : arg_descriptor < uint64_t > arg_store_quorum_history = {
" store-quorum-history " ,
" Store the service node quorum history for the last N blocks to allow historic quorum lookups "
" (e.g. by a block explorer). Specify the number of blocks of history to store, or 1 to store "
" the entire history. Requires considerably more memory and block chain storage. " ,
0 } ;
2019-10-15 19:45:58 +02:00
// Loads stubs that fail if invoked. The stubs are replaced in the cryptonote_protocol/quorumnet.cpp glue code.
2020-08-26 03:44:11 +02:00
[[noreturn]] static void need_core_init ( std : : string_view stub_name ) {
throw std : : logic_error ( " Internal error: core callback initialization was not performed for " s + std : : string ( stub_name ) ) ;
2019-10-15 19:45:58 +02:00
}
Replace epee http rpc server with uWebSockets
This replaces the NIH epee http server which does not work all that well
with an external C++ library called uWebSockets. Fundamentally this
gives the following advantages:
- Much less code to maintain
- Just one thread for handling HTTP connections versus epee's pool of
threads
- Uses existing LokiMQ job server and existing thread pool for handling
the actual tasks; they are processed/scheduled in the same "rpc" or
"admin" queues as lokimq rpc calls. One notable benefit is that "admin"
rpc commands get their own queue (and thus cannot be delayed by long rpc
commands). Currently the lokimq threads and the http rpc thread pool
and the p2p thread pool and the job queue thread pool and the dns lookup
thread pool and... are *all* different thread pools; this is a step
towards consolidating them.
- Very little mutex contention (which has been a major problem with epee
RPC in the past): there is one mutex (inside uWebSockets) for putting
responses back into the thread managing the connection; everything
internally gets handled through (lock-free) lokimq inproc sockets.
- Faster RPC performance on average, and much better worst case
performance. Epee's http interface seems to have some race condition
that ocassionally stalls a request (even a very simple one) for a dozen
or more seconds for no good reason.
- Long polling gets redone here to no longer need threads; instead we
just store the request and respond when the thread pool, or else in a
timer (that runs once/second) for timing out long polls.
---
The basic idea of how this works from a high level:
We launch a single thread to handle HTTP RPC requests and response data.
This uWebSockets thread is essentially running an event loop: it never
actually handles any logic; it only serves to shuttle data that arrives
in a request to some other thread, and then, at some later point, to
send some reply back to that waiting connection. Everything is
asynchronous and non-blocking here: the basic uWebSockets event loop
just operates as things arrive, passes it off immediately, and goes back
to waiting for the next thing to arrive.
The basic flow is like this:
0. uWS thread -- listens on localhost:22023
1. uWS thread -- incoming request on localhost:22023
2. uWS thread -- fires callback, which injects the task into the LokiMQ job queue
3. LMQ main loop -- schedules it as an RPC job
4. LMQ rpc thread -- Some LokiMQ thread runs it, gets the result
5. LMQ rpc thread -- Result gets queued up for the uWS thread
6. uWS thread -- takes the request and starts sending it
(asynchronously) back to the requestor.
In more detail:
uWebSockets has registered has registered handlers for non-jsonrpc
requests (legacy JSON or binary). If the port is restricted then admin
commands get mapped to a "Access denied" response handler, otherwise
public commands (and admin commands on an unrestricted port) go to the
rpc command handler.
POST requests to /json_rpc have their own handler; this is a little
different than the above because it has to parse the request before it
can determine whether it is allowed or not, but once this is done it
continues roughly the same as legacy/binary requests.
uWebSockets then listens on the given IP/port for new incoming requests,
and starts listening for requests in a thread (we own this thread).
When a request arrives, it fires the event handler for that request.
(This may happen multiple times, if the client is sending a bunch of
data in a POST request). Once we have the full request, we then queue
the job in LokiMQ, putting it in the "rpc" or "admin" command
categories. (The one practical different here is that "admin" is
configured to be allowed to start up its own thread if all other threads
are busy, while "rpc" commands are prioritized along with everything
else.) LokiMQ then schedules this, along with native LokiMQ "rpc." or
"admin." requests.
When a LMQ worker thread becomes available, the RPC command gets called
in it and runs. Whatever output it produces (or error message, if it
throws) then gets wrapped up in jsonrpc boilerplate (if necessary), and
delivered to the uWebSockets thread to be sent in reply to that request.
uWebSockets picks up the data and sends whatever it can without
blocking, then buffers whatever it couldn't send to be sent again in a
later event loop iteration once the requestor can accept more data.
(This part is outside lokid; we only have to give uWS the data and let
it worry about delivery).
---
PR specifics:
Things removed from this PR:
1. ssl settings; with this PR the HTTP RPC interface is plain-text. The
previous default generated a self-signed certificate for the server on
startup and then the client accepted any certificate. This is actually
*worse* than unencrypted because it is entirely MITM-readable and yet
might make people think that their RPC communication is encrypted, and
setting up actual certificates is difficult enough that I think most
people don't bother.
uWebSockets *does* support HTTPS, and we could glue the existing options
into it, but I'm not convinced it's worthwhile: it works much better to
put HTTPS in a front-end proxy holding the certificate that proxies
requests to the backend (which can then listen in restricted mode on
some localhost port). One reason this is better is that it is much
easier to reload and/or restart such a front-end server, while
certificate updates with lokid require a full restart. Another reason
is that you get an error page instead of a timeout if something is wrong
with the backend. Finally we also save having to generate a temporary
certificate on *every* lokid invocation.
2. HTTP Digest authentication. Digest authentication is obsolete (and
was already obsolete when it got added to Monero). HTTP-Digest was
originally an attempt to provide a password authentication mechanism
that does not leak the password in transit, but still required that the
server know the password. It only has marginal value against replay
attacks, and is made entirely obsolete by sending traffic over HTTPS
instead. No client out there supports Digest but *not* Basic auth, and
so given the limited usefulness it seems pointless to support more than
Basic auth for HTTP RPC login.
What's worse is that epee's HTTP Digest authentication is a terrible
implementation: it uses boost::spirit -- a recursive descent parser
meant for building complex language grammars -- just to parse a single
HTTP header for Digest auth. This is a big load of crap that should
never have been accepted upstream, and that we should get rid of (even
if we wanted to support Digest auth it takes less than 100 lines of code
to do it when *not* using a recursive descent parser).
2020-06-29 01:23:06 +02:00
2020-08-26 03:44:11 +02:00
void ( * long_poll_trigger ) ( tx_memory_pool & pool ) = [ ] ( tx_memory_pool & ) { need_core_init ( " long_poll_trigger " sv ) ; } ;
quorumnet_new_proc * quorumnet_new = [ ] ( core & ) - > void * { need_core_init ( " quorumnet_new " sv ) ; } ;
quorumnet_init_proc * quorumnet_init = [ ] ( core & , void * ) { need_core_init ( " quorumnet_init " sv ) ; } ;
quorumnet_delete_proc * quorumnet_delete = [ ] ( void * & ) { need_core_init ( " quorumnet_delete " sv ) ; } ;
quorumnet_relay_obligation_votes_proc * quorumnet_relay_obligation_votes = [ ] ( void * , const std : : vector < service_nodes : : quorum_vote_t > & ) { need_core_init ( " quorumnet_relay_obligation_votes " sv ) ; } ;
quorumnet_send_blink_proc * quorumnet_send_blink = [ ] ( core & , const std : : string & ) - > std : : future < std : : pair < blink_result , std : : string > > { need_core_init ( " quorumnet_send_blink " sv ) ; } ;
quorumnet_pulse_relay_message_to_quorum_proc * quorumnet_pulse_relay_message_to_quorum = [ ] ( void * , pulse : : message const & , service_nodes : : quorum const & , bool ) - > void { need_core_init ( " quorumnet_pulse_relay_message_to_quorum " sv ) ; } ;
2019-07-17 08:01:54 +02:00
2014-03-03 23:07:58 +01:00
//-----------------------------------------------------------------------------------------------
2020-04-23 07:52:19 +02:00
core : : core ( )
: m_mempool ( m_blockchain_storage )
, m_service_node_list ( m_blockchain_storage )
, m_blockchain_storage ( m_mempool , m_service_node_list )
, m_quorum_cop ( * this )
2020-05-25 08:08:05 +02:00
, m_miner ( this , [ this ] ( const cryptonote : : block & b , uint64_t height , unsigned int threads , crypto : : hash & hash ) {
2020-06-25 05:31:44 +02:00
hash = cryptonote : : get_block_longhash_w_blockchain ( m_nettype , & m_blockchain_storage , b , height , threads ) ;
2020-05-29 05:37:35 +02:00
return true ;
2020-05-25 08:08:05 +02:00
} )
2020-04-23 07:52:19 +02:00
, m_pprotocol ( & m_protocol_stub )
, m_starter_message_showed ( false )
, m_target_blockchain_height ( 0 )
, m_last_json_checkpoints_update ( 0 )
2022-05-16 22:55:05 +02:00
, m_nettype ( network_type : : UNDEFINED )
2020-04-23 07:52:19 +02:00
, m_last_storage_server_ping ( 0 )
2021-01-04 05:21:21 +01:00
, m_last_lokinet_ping ( 0 )
2020-04-23 07:52:19 +02:00
, m_pad_transactions ( false )
2021-04-20 08:22:00 +02:00
, ss_version { 0 }
, lokinet_version { 0 }
2014-03-03 23:07:58 +01:00
{
2017-07-02 19:42:45 +02:00
m_checkpoints_updating . clear ( ) ;
2014-03-03 23:07:58 +01:00
}
void core : : set_cryptonote_protocol ( i_cryptonote_protocol * pprotocol )
{
if ( pprotocol )
m_pprotocol = pprotocol ;
else
m_pprotocol = & m_protocol_stub ;
}
2014-09-25 07:15:28 +02:00
//-----------------------------------------------------------------------------------------------
2019-07-11 06:16:55 +02:00
bool core : : update_checkpoints_from_json_file ( )
2014-09-25 07:15:28 +02:00
{
2015-04-22 10:36:39 +02:00
if ( m_checkpoints_updating . test_and_set ( ) ) return true ;
2019-05-01 07:57:41 +02:00
// load json checkpoints every 10min and verify them with respect to what blocks we already have
2014-09-26 07:01:58 +02:00
bool res = true ;
2019-05-01 08:01:17 +02:00
if ( time ( NULL ) - m_last_json_checkpoints_update > = 600 )
2014-09-25 07:15:28 +02:00
{
2019-07-11 06:16:55 +02:00
res = m_blockchain_storage . update_checkpoints_from_json_file ( m_checkpoints_path ) ;
2014-09-29 22:30:47 +02:00
m_last_json_checkpoints_update = time ( NULL ) ;
}
2015-04-22 10:36:39 +02:00
m_checkpoints_updating . clear ( ) ;
2014-09-30 21:24:43 +02:00
// if anything fishy happened getting new checkpoints, bring down the house
if ( ! res )
{
graceful_exit ( ) ;
}
2014-09-26 07:01:58 +02:00
return res ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------
2015-02-05 10:11:20 +01:00
void core : : stop ( )
{
2018-11-30 15:53:38 +01:00
m_miner . stop ( ) ;
2016-12-04 13:27:45 +01:00
m_blockchain_storage . cancel ( ) ;
2015-02-05 10:11:20 +01:00
}
//-----------------------------------------------------------------------------------
2015-12-09 00:06:29 +01:00
void core : : init_options ( boost : : program_options : : options_description & desc )
2014-03-03 23:07:58 +01:00
{
2018-01-21 16:29:55 +01:00
command_line : : add_arg ( desc , arg_data_dir ) ;
2017-10-28 17:06:43 +02:00
command_line : : add_arg ( desc , arg_test_drop_download ) ;
command_line : : add_arg ( desc , arg_test_drop_download_height ) ;
command_line : : add_arg ( desc , arg_testnet_on ) ;
2020-08-11 23:41:54 +02:00
command_line : : add_arg ( desc , arg_devnet_on ) ;
2018-06-14 21:11:49 +02:00
command_line : : add_arg ( desc , arg_regtest_on ) ;
2019-11-01 16:16:37 +01:00
command_line : : add_arg ( desc , arg_keep_fakechain ) ;
2018-06-14 21:11:49 +02:00
command_line : : add_arg ( desc , arg_fixed_difficulty ) ;
2019-09-17 03:20:15 +02:00
command_line : : add_arg ( desc , arg_dev_allow_local ) ;
2017-10-28 17:06:43 +02:00
command_line : : add_arg ( desc , arg_prep_blocks_threads ) ;
command_line : : add_arg ( desc , arg_fast_block_sync ) ;
command_line : : add_arg ( desc , arg_show_time_stats ) ;
command_line : : add_arg ( desc , arg_block_sync_size ) ;
2017-11-30 16:35:52 +01:00
command_line : : add_arg ( desc , arg_offline ) ;
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
command_line : : add_arg ( desc , arg_block_download_max_size ) ;
2018-07-18 23:24:53 +02:00
command_line : : add_arg ( desc , arg_max_txpool_weight ) ;
2018-07-12 04:07:34 +02:00
command_line : : add_arg ( desc , arg_service_node ) ;
2019-06-04 06:15:16 +02:00
command_line : : add_arg ( desc , arg_public_ip ) ;
2019-10-15 19:45:58 +02:00
command_line : : add_arg ( desc , arg_storage_server_port ) ;
command_line : : add_arg ( desc , arg_quorumnet_port ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
2018-11-02 23:27:25 +01:00
command_line : : add_arg ( desc , arg_pad_transactions ) ;
2018-09-29 22:18:08 +02:00
command_line : : add_arg ( desc , arg_block_notify ) ;
2021-01-04 01:09:45 +01:00
#if 0 // TODO(oxen): Pruning not supported because of Service Node List
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
command_line : : add_arg ( desc , arg_prune_blockchain ) ;
2019-10-23 04:44:56 +02:00
# endif
2019-01-09 15:29:28 +01:00
command_line : : add_arg ( desc , arg_reorg_notify ) ;
2019-05-08 13:11:23 +02:00
command_line : : add_arg ( desc , arg_keep_alt_blocks ) ;
2017-02-09 22:29:35 +01:00
2019-07-17 08:01:54 +02:00
command_line : : add_arg ( desc , arg_store_quorum_history ) ;
2021-04-08 18:38:51 +02:00
command_line : : add_arg ( desc , arg_omq_quorumnet_public ) ;
2018-11-15 07:17:51 +01:00
2017-02-21 16:06:07 +01:00
miner : : init_options ( desc ) ;
2017-08-29 12:14:42 +02:00
BlockchainDB : : init_options ( desc ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-01-29 23:10:53 +01:00
bool core : : handle_command_line ( const boost : : program_options : : variables_map & vm )
2014-03-03 23:07:58 +01:00
{
2022-05-16 22:55:05 +02:00
if ( m_nettype ! = network_type : : FAKECHAIN )
2018-02-16 12:04:04 +01:00
{
const bool testnet = command_line : : get_arg ( vm , arg_testnet_on ) ;
2020-08-11 23:41:54 +02:00
const bool devnet = command_line : : get_arg ( vm , arg_devnet_on ) ;
2022-05-16 22:55:05 +02:00
m_nettype = testnet ? network_type : : TESTNET : devnet ? network_type : : DEVNET : network_type : : MAINNET ;
2018-02-16 12:04:04 +01:00
}
2021-03-25 16:44:30 +01:00
m_check_uptime_proof_interval . interval ( get_net_config ( ) . UPTIME_PROOF_CHECK_INTERVAL ) ;
2015-01-29 23:10:53 +01:00
2020-10-22 19:55:33 +02:00
m_config_folder = fs : : u8path ( command_line : : get_arg ( vm , arg_data_dir ) ) ;
2015-01-29 23:10:53 +01:00
2017-10-28 17:06:43 +02:00
test_drop_download_height ( command_line : : get_arg ( vm , arg_test_drop_download_height ) ) ;
2018-11-02 23:27:25 +01:00
m_pad_transactions = get_arg ( vm , arg_pad_transactions ) ;
2017-11-30 16:35:52 +01:00
m_offline = get_arg ( vm , arg_offline ) ;
2017-10-28 17:06:43 +02:00
if ( command_line : : get_arg ( vm , arg_test_drop_download ) = = true )
2015-12-14 05:54:39 +01:00
test_drop_download ( ) ;
2018-07-12 04:07:34 +02:00
2019-09-17 03:20:15 +02:00
if ( command_line : : get_arg ( vm , arg_dev_allow_local ) )
m_service_node_list . debug_allow_local_ips = true ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
m_service_node = command_line : : get_arg ( vm , arg_service_node ) ;
2019-10-07 17:16:20 +02:00
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( m_service_node ) {
2019-06-04 06:15:16 +02:00
/// TODO: parse these options early, before we start p2p server etc?
2019-10-15 19:45:58 +02:00
m_quorumnet_port = command_line : : get_arg ( vm , arg_quorumnet_port ) ;
2019-06-04 06:15:16 +02:00
2021-04-08 18:38:51 +02:00
bool args_okay = true ;
2019-10-15 19:45:58 +02:00
if ( m_quorumnet_port = = 0 ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Quorumnet port cannot be 0; please specify a valid port to listen on with: '--{} <port>' " , arg_quorumnet_port . name ) ;
2021-04-08 18:38:51 +02:00
args_okay = false ;
2019-06-04 06:15:16 +02:00
}
const std : : string pub_ip = command_line : : get_arg ( vm , arg_public_ip ) ;
2019-06-06 09:39:54 +02:00
if ( pub_ip . size ( ) )
{
if ( ! epee : : string_tools : : get_ip_int32_from_string ( m_sn_public_ip , pub_ip ) ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Unable to parse IPv4 public address from: {} " , pub_ip ) ;
2021-04-08 18:38:51 +02:00
args_okay = false ;
2019-06-06 09:39:54 +02:00
}
2019-06-25 18:31:38 +02:00
if ( ! epee : : net_utils : : is_ip_public ( m_sn_public_ip ) ) {
2019-09-17 03:20:15 +02:00
if ( m_service_node_list . debug_allow_local_ips ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , " Address given for public-ip is not public; allowing it because dev-allow-local-ips was specified. This service node WILL NOT WORK ON THE PUBLIC OXEN NETWORK! " ) ;
2019-09-17 03:20:15 +02:00
} else {
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Address given for public-ip is not public: {} " , epee : : string_tools : : get_ip_string_from_int32 ( m_sn_public_ip ) ) ;
2021-04-08 18:38:51 +02:00
args_okay = false ;
2019-09-17 03:20:15 +02:00
}
2019-06-06 09:39:54 +02:00
}
}
else
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Please specify an IPv4 public address which the service node & storage server is accessible from with: '--{} <ip address>' " , arg_public_ip . name ) ;
2021-04-08 18:38:51 +02:00
args_okay = false ;
2019-06-04 06:15:16 +02:00
}
2021-04-08 18:38:51 +02:00
if ( ! args_okay ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat ,
" IMPORTANT: One or more required service node-related configuration settings/options were omitted or invalid please fix them and restart oxend. " ) ;
2019-06-04 06:15:16 +02:00
return false ;
}
}
2014-03-03 23:07:58 +01:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
uint64_t core : : get_current_blockchain_height ( ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_current_blockchain_height ( ) ;
}
//-----------------------------------------------------------------------------------------------
2021-08-05 02:27:32 +02:00
std : : pair < uint64_t , crypto : : hash > core : : get_blockchain_top ( ) const
2014-03-03 23:07:58 +01:00
{
2021-08-05 02:27:32 +02:00
std : : pair < uint64_t , crypto : : hash > result ;
result . second = m_blockchain_storage . get_tail_id ( result . first ) ;
return result ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : get_blocks ( uint64_t start_offset , size_t count , std : : vector < std : : pair < std : : string , block > > & blocks , std : : vector < std : : string > & txs ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_blocks ( start_offset , count , blocks , txs ) ;
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : get_blocks ( uint64_t start_offset , size_t count , std : : vector < std : : pair < std : : string , block > > & blocks ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_blocks ( start_offset , count , blocks ) ;
2017-01-15 17:05:55 +01:00
}
//-----------------------------------------------------------------------------------------------
2018-04-16 01:16:02 +02:00
bool core : : get_blocks ( uint64_t start_offset , size_t count , std : : vector < block > & blocks ) const
2017-01-15 17:05:55 +01:00
{
2020-09-02 06:21:44 +02:00
return m_blockchain_storage . get_blocks_only ( start_offset , count , blocks ) ;
2017-01-15 17:05:55 +01:00
}
//-----------------------------------------------------------------------------------------------
2022-07-12 02:40:50 +02:00
bool core : : get_blocks ( const std : : vector < crypto : : hash > & block_ids , std : : vector < std : : pair < std : : string , block > > blocks , std : : unordered_set < crypto : : hash > * missed_bs ) const
2021-08-19 05:43:26 +02:00
{
return m_blockchain_storage . get_blocks ( block_ids , blocks , missed_bs ) ;
}
//-----------------------------------------------------------------------------------------------
2022-07-12 02:40:50 +02:00
bool core : : get_transactions ( const std : : vector < crypto : : hash > & txs_ids , std : : vector < std : : string > & txs , std : : unordered_set < crypto : : hash > * missed_txs ) const
2017-01-15 17:05:55 +01:00
{
return m_blockchain_storage . get_transactions_blobs ( txs_ids , txs , missed_txs ) ;
}
//-----------------------------------------------------------------------------------------------
2022-07-12 02:40:50 +02:00
bool core : : get_split_transactions_blobs ( const std : : vector < crypto : : hash > & txs_ids , std : : vector < std : : tuple < crypto : : hash , std : : string , crypto : : hash , std : : string > > & txs , std : : unordered_set < crypto : : hash > * missed_txs ) const
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
{
return m_blockchain_storage . get_split_transactions_blobs ( txs_ids , txs , missed_txs ) ;
}
//-----------------------------------------------------------------------------------------------
2021-08-19 05:43:26 +02:00
bool core : : get_transactions ( const std : : vector < crypto : : hash > & txs_ids , std : : vector < transaction > & txs , std : : unordered_set < crypto : : hash > * missed_txs ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_transactions ( txs_ids , txs , missed_txs ) ;
}
//-----------------------------------------------------------------------------------------------
2018-04-16 01:16:02 +02:00
bool core : : get_alternative_blocks ( std : : vector < block > & blocks ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_alternative_blocks ( blocks ) ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
size_t core : : get_alternative_blocks_count ( ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_alternative_blocks_count ( ) ;
}
2020-01-19 23:44:18 +01:00
static std : : string time_ago_str ( time_t now , time_t then ) {
if ( then > = now )
return " now " s ;
if ( then = = 0 )
return " never " s ;
int seconds = now - then ;
if ( seconds > = 60 )
return std : : to_string ( seconds / 60 ) + " m " + std : : to_string ( seconds % 60 ) + " s " ;
return std : : to_string ( seconds % 60 ) + " s " ;
}
2020-12-01 01:56:43 +01:00
// Returns a bool on whether the service node is currently active
bool core : : is_active_sn ( ) const
{
auto info = get_my_sn_info ( ) ;
return ( info & & info - > is_active ( ) ) ;
}
// Returns the service nodes info
std : : shared_ptr < const service_nodes : : service_node_info > core : : get_my_sn_info ( ) const
{
auto & snl = get_service_node_list ( ) ;
const auto & pubkey = get_service_keys ( ) . pub ;
auto states = snl . get_service_node_list_state ( { pubkey } ) ;
if ( states . empty ( ) )
return nullptr ;
else
{
return states [ 0 ] . info ;
}
}
2020-01-19 23:44:18 +01:00
// Returns a string for systemd status notifications such as:
2021-01-04 05:21:21 +01:00
// Height: 1234567, SN: active, proof: 55m12s, storage: 4m48s, lokinet: 47s
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
std : : string core : : get_status_string ( ) const
2020-01-19 23:44:18 +01:00
{
std : : string s ;
s . reserve ( 128 ) ;
2021-01-04 04:19:42 +01:00
s + = ' v ' ; s + = OXEN_VERSION_STR ;
2020-03-18 09:28:06 +01:00
s + = " ; Height: " ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
s + = std : : to_string ( get_blockchain_storage ( ) . get_current_blockchain_height ( ) ) ;
2020-01-19 23:44:18 +01:00
s + = " , SN: " ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( ! service_node ( ) )
2020-01-19 23:44:18 +01:00
s + = " no " ;
else
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
auto & snl = get_service_node_list ( ) ;
const auto & pubkey = get_service_keys ( ) . pub ;
auto states = snl . get_service_node_list_state ( { pubkey } ) ;
2020-01-19 23:44:18 +01:00
if ( states . empty ( ) )
s + = " not registered " ;
else
{
auto & info = * states [ 0 ] . info ;
if ( ! info . is_fully_funded ( ) )
s + = " awaiting contr. " ;
else if ( info . is_active ( ) )
s + = " active " ;
else if ( info . is_decommissioned ( ) )
s + = " decomm. " ;
uint64_t last_proof = 0 ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
snl . access_proof ( pubkey , [ & ] ( auto & proof ) { last_proof = proof . timestamp ; } ) ;
2020-01-19 23:44:18 +01:00
s + = " , proof: " ;
time_t now = std : : time ( nullptr ) ;
s + = time_ago_str ( now , last_proof ) ;
s + = " , storage: " ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
s + = time_ago_str ( now , m_last_storage_server_ping ) ;
2021-01-04 05:21:21 +01:00
s + = " , lokinet: " ;
s + = time_ago_str ( now , m_last_lokinet_ping ) ;
2020-01-19 23:44:18 +01:00
}
}
return s ;
}
2014-03-03 23:07:58 +01:00
//-----------------------------------------------------------------------------------------------
2018-11-23 14:47:51 +01:00
bool core : : init ( const boost : : program_options : : variables_map & vm , const cryptonote : : test_options * test_options , const GetCheckpointsCallback & get_checkpoints /* = nullptr */ )
2014-03-03 23:07:58 +01:00
{
2017-01-09 00:50:29 +01:00
start_time = std : : time ( nullptr ) ;
2018-06-14 21:11:49 +02:00
const bool regtest = command_line : : get_arg ( vm , arg_regtest_on ) ;
if ( test_options ! = NULL | | regtest )
2018-02-16 12:04:04 +01:00
{
2022-05-16 22:55:05 +02:00
m_nettype = network_type : : FAKECHAIN ;
2018-02-16 12:04:04 +01:00
}
2018-11-21 00:43:09 +01:00
2015-01-29 23:10:53 +01:00
bool r = handle_command_line ( vm ) ;
2019-06-04 06:15:16 +02:00
/// Currently terminating before blockchain is initialized results in a crash
/// during deinitialization... TODO: fix that
CHECK_AND_ASSERT_MES ( r , false , " Failed to apply command line options. " ) ;
2017-01-28 01:36:39 +01:00
2017-08-29 12:14:42 +02:00
std : : string db_sync_mode = command_line : : get_arg ( vm , cryptonote : : arg_db_sync_mode ) ;
bool db_salvage = command_line : : get_arg ( vm , cryptonote : : arg_db_salvage ) ! = 0 ;
2017-10-28 17:06:43 +02:00
bool fast_sync = command_line : : get_arg ( vm , arg_fast_block_sync ) ! = 0 ;
uint64_t blocks_threads = command_line : : get_arg ( vm , arg_prep_blocks_threads ) ;
2018-07-18 23:24:53 +02:00
size_t max_txpool_weight = command_line : : get_arg ( vm , arg_max_txpool_weight ) ;
2019-12-04 06:27:51 +01:00
bool const prune_blockchain = false ; /* command_line::get_arg(vm, arg_prune_blockchain); */
2019-05-08 13:11:23 +02:00
bool keep_alt_blocks = command_line : : get_arg ( vm , arg_keep_alt_blocks ) ;
2019-11-01 16:16:37 +01:00
bool keep_fakechain = command_line : : get_arg ( vm , arg_keep_fakechain ) ;
2015-03-25 16:41:30 +01:00
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
r = init_service_keys ( ) ;
CHECK_AND_ASSERT_MES ( r , false , " Failed to create or load service keys " ) ;
if ( m_service_node )
2018-07-12 04:07:34 +02:00
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
// Only use our service keys for our service node if we are running in SN mode:
m_service_node_list . set_my_service_node_keys ( & m_service_keys ) ;
2018-07-12 04:07:34 +02:00
}
2020-10-22 19:55:33 +02:00
auto folder = m_config_folder ;
2022-05-16 22:55:05 +02:00
if ( m_nettype = = network_type : : FAKECHAIN )
2015-12-31 18:09:00 +01:00
folder / = " fake " ;
2016-02-04 19:09:45 +01:00
// make sure the data directory exists, and try to lock it
2020-11-12 20:16:25 +01:00
if ( std : : error_code ec ; ! fs : : is_directory ( folder , ec ) & & ! fs : : create_directories ( folder , ec ) & & ec )
2020-10-22 19:55:33 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to create directory " + folder . u8string ( ) + ( ec ? " : " + ec . message ( ) : " " s ) ) ;
2020-10-22 19:55:33 +02:00
return false ;
}
2016-02-04 19:09:45 +01:00
2019-06-26 16:39:41 +02:00
std : : unique_ptr < BlockchainDB > db ( new_db ( ) ) ;
2020-10-22 19:55:33 +02:00
if ( ! db )
2015-03-25 16:41:30 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to initialize a database " ) ;
2015-03-25 16:41:30 +01:00
return false ;
}
2021-02-12 05:19:30 +01:00
auto ons_db_file_path = folder / " ons.db " ;
2021-04-15 07:47:45 +02:00
if ( fs : : exists ( folder / " lns.db " ) )
2021-04-14 09:25:53 +02:00
ons_db_file_path = folder / " lns.db " ;
2022-04-29 01:51:14 +02:00
auto sqlite_db_file_path = folder / " sqlite.db " ;
2022-05-16 22:55:05 +02:00
if ( m_nettype = = network_type : : FAKECHAIN )
2022-04-29 01:51:14 +02:00
{
sqlite_db_file_path = " :memory: " ;
}
auto sqliteDB = std : : make_shared < cryptonote : : BlockchainSQLite > ( m_nettype , sqlite_db_file_path ) ;
2020-09-10 02:02:40 +02:00
2015-03-25 16:41:30 +01:00
folder / = db - > get_db_name ( ) ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Loading blockchain from folder {} ... " , folder ) ;
2015-03-25 16:41:30 +01:00
2018-08-01 12:09:35 +02:00
// default to fast:async:1 if overridden
2017-08-19 20:36:51 +02:00
blockchain_db_sync_mode sync_mode = db_defaultsync ;
2018-08-01 12:09:35 +02:00
bool sync_on_blocks = true ;
uint64_t sync_threshold = 1 ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
2022-05-16 22:55:05 +02:00
if ( m_nettype = = network_type : : FAKECHAIN & & ! keep_fakechain )
2018-06-14 21:11:49 +02:00
{
// reset the db by removing the database file before opening it
2020-10-22 19:55:33 +02:00
if ( ! db - > remove_data_file ( folder ) )
2018-06-14 21:11:49 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to remove data file in {} " , folder ) ;
2018-06-14 21:11:49 +02:00
return false ;
}
2021-02-12 05:19:30 +01:00
fs : : remove ( ons_db_file_path ) ;
2018-06-14 21:11:49 +02:00
}
2015-03-25 16:41:30 +01:00
try
{
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
uint64_t db_flags = 0 ;
std : : vector < std : : string > options ;
boost : : trim ( db_sync_mode ) ;
boost : : split ( options , db_sync_mode , boost : : is_any_of ( " : " ) ) ;
2018-05-28 13:12:19 +02:00
const bool db_sync_mode_is_default = command_line : : is_arg_defaulted ( vm , cryptonote : : arg_db_sync_mode ) ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
for ( const auto & option : options )
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " option: {} " , option ) ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
2017-01-01 04:38:55 +01:00
// default to fast:async:1
2017-08-19 16:27:13 +02:00
uint64_t DEFAULT_FLAGS = DBF_FAST ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
if ( options . size ( ) = = 0 )
{
2017-01-01 04:38:55 +01:00
// default to fast:async:1
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
db_flags = DEFAULT_FLAGS ;
}
bool safemode = false ;
if ( options . size ( ) > = 1 )
{
if ( options [ 0 ] = = " safe " )
{
safemode = true ;
2017-08-19 16:27:13 +02:00
db_flags = DBF_SAFE ;
2018-05-28 13:12:19 +02:00
sync_mode = db_sync_mode_is_default ? db_defaultsync : db_nosync ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
}
else if ( options [ 0 ] = = " fast " )
2017-08-19 20:36:51 +02:00
{
2017-08-19 16:27:13 +02:00
db_flags = DBF_FAST ;
2018-05-28 13:12:19 +02:00
sync_mode = db_sync_mode_is_default ? db_defaultsync : db_async ;
2017-08-19 20:36:51 +02:00
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
else if ( options [ 0 ] = = " fastest " )
2017-01-01 04:38:55 +01:00
{
2017-08-19 16:27:13 +02:00
db_flags = DBF_FASTEST ;
2018-08-01 12:09:35 +02:00
sync_threshold = 1000 ; // default to fastest:async:1000
2018-05-28 13:12:19 +02:00
sync_mode = db_sync_mode_is_default ? db_defaultsync : db_async ;
2017-01-01 04:38:55 +01:00
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
else
db_flags = DEFAULT_FLAGS ;
}
if ( options . size ( ) > = 2 & & ! safemode )
{
if ( options [ 1 ] = = " sync " )
2018-05-28 13:12:19 +02:00
sync_mode = db_sync_mode_is_default ? db_defaultsync : db_sync ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
else if ( options [ 1 ] = = " async " )
2018-05-28 13:12:19 +02:00
sync_mode = db_sync_mode_is_default ? db_defaultsync : db_async ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
}
if ( options . size ( ) > = 3 & & ! safemode )
{
2016-08-29 17:42:52 +02:00
char * endptr ;
2018-08-01 12:09:35 +02:00
uint64_t threshold = strtoull ( options [ 2 ] . c_str ( ) , & endptr , 0 ) ;
if ( * endptr = = ' \0 ' | | ! strcmp ( endptr , " blocks " ) )
{
sync_on_blocks = true ;
sync_threshold = threshold ;
}
else if ( ! strcmp ( endptr , " bytes " ) )
{
sync_on_blocks = false ;
sync_threshold = threshold ;
}
else
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Invalid db sync mode: {} " , options [ 2 ] ) ;
2018-08-01 12:09:35 +02:00
return false ;
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
}
2017-07-31 00:59:52 +02:00
if ( db_salvage )
2017-08-19 16:27:13 +02:00
db_flags | = DBF_SALVAGE ;
2017-07-31 00:59:52 +02:00
2020-10-22 19:55:33 +02:00
db - > open ( folder , m_nettype , db_flags ) ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
if ( ! db - > m_open )
2015-12-14 05:54:39 +01:00
return false ;
2015-03-25 16:41:30 +01:00
}
catch ( const DB_ERROR & e )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Error opening database: {} " , e . what ( ) ) ;
2015-03-25 16:41:30 +01:00
return false ;
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
m_blockchain_storage . set_user_options ( blocks_threads ,
2018-08-01 12:09:35 +02:00
sync_on_blocks , sync_threshold , sync_mode , fast_sync ) ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
2022-07-27 22:02:12 +02:00
// We need this hook to get added before the block hook below, so that it fires first and
// catches the start of a reorg before the block hook fires for the block in the reorg.
2019-01-09 15:29:28 +01:00
try
{
if ( ! command_line : : is_arg_defaulted ( vm , arg_reorg_notify ) )
2022-07-27 22:02:12 +02:00
m_blockchain_storage . hook_block_post_add (
[ this , notify = tools : : Notify ( command_line : : get_arg ( vm , arg_reorg_notify ) ) ]
( const auto & info ) {
if ( ! info . reorg )
return ;
auto h = get_current_blockchain_height ( ) ;
notify . notify (
" %s " , info . split_height ,
" %h " , h ,
" %n " , h - info . split_height ) ;
} ) ;
2019-01-09 15:29:28 +01:00
}
catch ( const std : : exception & e )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to parse reorg notify spec " ) ;
2019-01-09 15:29:28 +01:00
}
2019-02-06 18:09:28 +01:00
try
{
2022-07-27 22:02:12 +02:00
if ( ! command_line : : is_arg_defaulted ( vm , arg_block_notify ) )
m_blockchain_storage . hook_block_post_add (
[ notify = tools : : Notify ( command_line : : get_arg ( vm , arg_block_notify ) ) ]
( const auto & info ) {
notify . notify ( " %s " , tools : : type_to_hex ( get_block_hash ( info . block ) ) ) ;
} ) ;
2019-02-06 18:09:28 +01:00
}
catch ( const std : : exception & e )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to parse block rate notify spec " ) ;
2019-02-06 18:09:28 +01:00
}
2021-06-18 01:06:51 +02:00
cryptonote : : test_options regtest_test_options { } ;
for ( auto [ it , end ] = get_hard_forks ( network_type : : MAINNET ) ;
it ! = end ;
it + + ) {
regtest_test_options . hard_forks . push_back ( hard_fork {
it - > version , it - > snode_revision , regtest_test_options . hard_forks . size ( ) , std : : time ( nullptr ) } ) ;
}
2018-09-19 06:05:07 +02:00
2019-05-01 07:57:41 +02:00
// Service Nodes
2022-07-27 21:58:13 +02:00
m_service_node_list . set_quorum_history_storage ( command_line : : get_arg ( vm , arg_store_quorum_history ) ) ;
// NOTE: Implicit dependency. Service node list needs to be hooked before checkpoints.
m_blockchain_storage . hook_blockchain_detached ( [ this ] ( const auto & info ) { m_service_node_list . blockchain_detached ( info . height ) ; } ) ;
m_blockchain_storage . hook_init ( [ this ] { m_service_node_list . init ( ) ; } ) ;
m_blockchain_storage . hook_validate_miner_tx ( [ this ] ( const auto & info ) { m_service_node_list . validate_miner_tx ( info ) ; } ) ;
m_blockchain_storage . hook_alt_block_add ( [ this ] ( const auto & info ) { m_service_node_list . alt_block_add ( info ) ; } ) ;
// NOTE: There is an implicit dependency on service node lists being hooked first!
m_blockchain_storage . hook_init ( [ this ] { m_quorum_cop . init ( ) ; } ) ;
m_blockchain_storage . hook_block_add ( [ this ] ( const auto & info ) { m_quorum_cop . block_add ( info . block , info . txs ) ; } ) ;
m_blockchain_storage . hook_blockchain_detached ( [ this ] ( const auto & info ) { m_quorum_cop . blockchain_detached ( info . height , info . by_pop_blocks ) ; } ) ;
m_blockchain_storage . hook_block_post_add ( [ this ] ( const auto & ) { update_omq_sns ( ) ; } ) ;
2019-05-01 07:57:41 +02:00
// Checkpoints
2020-10-22 19:55:33 +02:00
m_checkpoints_path = m_config_folder / fs : : u8path ( JSON_HASH_FILE_NAME ) ;
2019-05-01 07:57:41 +02:00
2021-02-12 05:19:30 +01:00
sqlite3 * ons_db = ons : : init_oxen_name_system ( ons_db_file_path , db - > is_read_only ( ) ) ;
if ( ! ons_db ) return false ;
2019-12-04 21:50:28 +01:00
2022-04-29 01:51:14 +02:00
2021-01-14 21:28:50 +01:00
init_oxenmq ( vm ) ;
2020-05-11 21:39:20 +02:00
2019-11-01 01:58:48 +01:00
const difficulty_type fixed_difficulty = command_line : : get_arg ( vm , arg_fixed_difficulty ) ;
2022-04-29 01:51:14 +02:00
r = m_blockchain_storage . init ( db . release ( ) , ons_db , std : : move ( sqliteDB ) , m_nettype , m_offline , regtest ? & regtest_test_options : test_options , fixed_difficulty , get_checkpoints ) ;
2019-11-28 02:11:46 +01:00
CHECK_AND_ASSERT_MES ( r , false , " Failed to initialize blockchain storage " ) ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
2018-07-18 23:24:53 +02:00
r = m_mempool . init ( max_txpool_weight ) ;
2017-05-14 15:06:55 +02:00
CHECK_AND_ASSERT_MES ( r , false , " Failed to initialize memory pool " ) ;
2016-01-29 16:09:17 +01:00
// now that we have a valid m_blockchain_storage, we can clean out any
// transactions in the pool that do not conform to the current fork
2021-06-18 01:06:51 +02:00
m_mempool . validate ( m_blockchain_storage . get_network_version ( ) ) ;
2016-01-29 16:09:17 +01:00
2017-10-28 17:06:43 +02:00
bool show_time_stats = command_line : : get_arg ( vm , arg_show_time_stats ) ! = 0 ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
m_blockchain_storage . set_show_time_stats ( show_time_stats ) ;
2014-03-03 23:07:58 +01:00
2017-10-28 17:06:43 +02:00
block_sync_size = command_line : : get_arg ( vm , arg_block_sync_size ) ;
2019-04-23 21:32:27 +02:00
if ( block_sync_size > BLOCKS_SYNCHRONIZING_MAX_COUNT )
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Error --block-sync-size cannot be greater than {} " , BLOCKS_SYNCHRONIZING_MAX_COUNT ) ;
2016-09-24 18:00:19 +02:00
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Loading checkpoints " ) ;
2019-07-11 06:16:55 +02:00
CHECK_AND_ASSERT_MES ( update_checkpoints_from_json_file ( ) , false , " One or more checkpoints loaded from json conflicted with existing checkpoints. " ) ;
2014-09-27 23:46:12 +02:00
2018-02-16 12:04:04 +01:00
r = m_miner . init ( vm , m_nettype ) ;
2016-03-25 07:22:06 +01:00
CHECK_AND_ASSERT_MES ( r , false , " Failed to initialize miner instance " ) ;
2014-03-03 23:07:58 +01:00
2019-05-08 13:11:23 +02:00
if ( ! keep_alt_blocks & & ! m_blockchain_storage . get_db ( ) . is_read_only ( ) )
m_blockchain_storage . get_db ( ) . drop_alt_blocks ( ) ;
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
if ( prune_blockchain )
{
// display a message if the blockchain is not pruned yet
2019-06-17 18:26:09 +02:00
if ( ! m_blockchain_storage . get_blockchain_pruning_seed ( ) )
2019-05-28 11:09:30 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Pruning blockchain... " ) ;
2019-05-28 11:09:30 +02:00
CHECK_AND_ASSERT_MES ( m_blockchain_storage . prune_blockchain ( ) , false , " Failed to prune blockchain " ) ;
}
else
{
CHECK_AND_ASSERT_MES ( m_blockchain_storage . update_blockchain_pruning ( ) , false , " Failed to update blockchain pruning " ) ;
}
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
}
2019-09-17 04:57:15 +02:00
return true ;
2014-03-03 23:07:58 +01:00
}
2019-10-07 17:16:20 +02:00
/// Loads a key pair from disk, if it exists, otherwise generates a new key pair and saves it to
/// disk.
///
/// get_pubkey - a function taking (privkey &, pubkey &) that sets the pubkey from the privkey;
/// returns true for success/false for failure
2020-08-29 21:56:19 +02:00
/// generate_pair - a void function taking (privkey &, pubkey &) that sets them to the generated values; can throw on error.
2019-10-07 17:16:20 +02:00
template < typename Privkey , typename Pubkey , typename GetPubkey , typename GeneratePair >
2020-10-22 19:55:33 +02:00
bool init_key ( const fs : : path & keypath , Privkey & privkey , Pubkey & pubkey , GetPubkey get_pubkey , GeneratePair generate_pair ) {
std : : error_code ec ;
if ( fs : : exists ( keypath , ec ) )
2018-07-12 04:07:34 +02:00
{
std : : string keystr ;
2020-10-22 19:55:33 +02:00
bool r = tools : : slurp_file ( keypath , keystr ) ;
2019-10-07 17:16:20 +02:00
memcpy ( & unwrap ( unwrap ( privkey ) ) , keystr . data ( ) , sizeof ( privkey ) ) ;
memwipe ( & keystr [ 0 ] , keystr . size ( ) ) ;
2020-10-22 19:55:33 +02:00
CHECK_AND_ASSERT_MES ( r , false , " failed to load service node key from " + keypath . u8string ( ) ) ;
2019-10-07 17:16:20 +02:00
CHECK_AND_ASSERT_MES ( keystr . size ( ) = = sizeof ( privkey ) , false ,
2020-10-22 19:55:33 +02:00
" service node key file " + keypath . u8string ( ) + " has an invalid size " ) ;
2018-07-12 04:07:34 +02:00
2019-10-07 17:16:20 +02:00
r = get_pubkey ( privkey , pubkey ) ;
2018-07-12 04:07:34 +02:00
CHECK_AND_ASSERT_MES ( r , false , " failed to generate pubkey from secret key " ) ;
}
else
{
2020-08-29 21:56:19 +02:00
try {
generate_pair ( privkey , pubkey ) ;
} catch ( const std : : exception & e ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " failed to generate keypair {} " , e . what ( ) ) ;
2020-08-29 21:56:19 +02:00
return false ;
}
2018-07-12 04:07:34 +02:00
2020-10-22 19:55:33 +02:00
bool r = tools : : dump_file ( keypath , tools : : view_guts ( privkey ) ) ;
CHECK_AND_ASSERT_MES ( r , false , " failed to save service node key to " + keypath . u8string ( ) ) ;
2018-07-12 04:07:34 +02:00
2020-10-22 19:55:33 +02:00
fs : : permissions ( keypath , fs : : perms : : owner_read , ec ) ;
2018-07-12 04:07:34 +02:00
}
2019-10-07 17:16:20 +02:00
return true ;
}
2018-07-12 04:07:34 +02:00
2019-10-07 17:16:20 +02:00
//-----------------------------------------------------------------------------------------------
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
bool core : : init_service_keys ( )
2019-10-07 17:16:20 +02:00
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
auto & keys = m_service_keys ;
2018-07-12 04:07:34 +02:00
2019-10-07 19:23:16 +02:00
static_assert (
sizeof ( crypto : : ed25519_public_key ) = = crypto_sign_ed25519_PUBLICKEYBYTES & &
sizeof ( crypto : : ed25519_secret_key ) = = crypto_sign_ed25519_SECRETKEYBYTES & &
sizeof ( crypto : : ed25519_signature ) = = crypto_sign_BYTES & &
sizeof ( crypto : : x25519_public_key ) = = crypto_scalarmult_curve25519_BYTES & &
sizeof ( crypto : : x25519_secret_key ) = = crypto_scalarmult_curve25519_BYTES ,
" Invalid ed25519/x25519 sizes " ) ;
2020-08-29 21:56:19 +02:00
// <data>/key_ed25519: Standard ed25519 secret key. We always have this, and generate one if it
// doesn't exist.
//
// As of Loki 8.x, if this exists and `key` doesn't, we use this key for everything. For
// compatibility with earlier versions we also allow `key` to contain a separate monero privkey
// for the SN keypair. (The main difference is that the Monero keypair is unclamped and that it
// only contains the private key value but not the secret key value that we need for full
// Ed25519 signing).
2019-10-07 19:23:16 +02:00
//
2020-10-22 19:55:33 +02:00
if ( ! init_key ( m_config_folder / " key_ed25519 " , keys . key_ed25519 , keys . pub_ed25519 ,
2019-10-07 19:23:16 +02:00
[ ] ( crypto : : ed25519_secret_key & sk , crypto : : ed25519_public_key & pk ) { crypto_sign_ed25519_sk_to_pk ( pk . data , sk . data ) ; return true ; } ,
[ ] ( crypto : : ed25519_secret_key & sk , crypto : : ed25519_public_key & pk ) { crypto_sign_ed25519_keypair ( pk . data , sk . data ) ; } )
)
return false ;
// Standard x25519 keys generated from the ed25519 keypair, used for encrypted communication between SNs
int rc = crypto_sign_ed25519_pk_to_curve25519 ( keys . pub_x25519 . data , keys . pub_ed25519 . data ) ;
CHECK_AND_ASSERT_MES ( rc = = 0 , false , " failed to convert ed25519 pubkey to x25519 " ) ;
crypto_sign_ed25519_sk_to_curve25519 ( keys . key_x25519 . data , keys . key_ed25519 . data ) ;
2020-08-29 21:56:19 +02:00
// Legacy primary SN key file; we only load this if it exists, otherwise we use `key_ed25519`
// for the primary SN keypair. (This key predates the Ed25519 keys and so is needed for
// backwards compatibility with existing active service nodes.) The legacy key consists of
// *just* the private point, but not the seed, and so cannot be used for full Ed25519 signatures
// (which rely on the seed for signing).
if ( m_service_node ) {
2020-10-22 19:55:33 +02:00
if ( std : : error_code ec ; ! fs : : exists ( m_config_folder / " key " , ec ) ) {
2020-08-29 21:56:19 +02:00
epee : : wipeable_string privkey_signhash ;
2020-08-31 06:08:47 +02:00
privkey_signhash . resize ( crypto_hash_sha512_BYTES ) ;
2020-08-29 21:56:19 +02:00
unsigned char * pk_sh_data = reinterpret_cast < unsigned char * > ( privkey_signhash . data ( ) ) ;
2020-08-31 06:08:47 +02:00
crypto_hash_sha512 ( pk_sh_data , keys . key_ed25519 . data , 32 /* first 32 bytes are the seed to be SHA512 hashed (the last 32 are just the pubkey) */ ) ;
// Clamp private key (as libsodium does and expects -- see https://www.jcraige.com/an-explainer-on-ed25519-clamping if you want the broader reasons)
2020-08-29 21:56:19 +02:00
pk_sh_data [ 0 ] & = 248 ;
2020-08-31 06:08:47 +02:00
pk_sh_data [ 31 ] & = 63 ; // (some implementations put 127 here, but with the |64 in the next line it is the same thing)
2020-08-29 21:56:19 +02:00
pk_sh_data [ 31 ] | = 64 ;
2020-09-02 16:39:25 +02:00
// Monero crypto requires a pointless check that the secret key is < basepoint, so calculate
2020-09-03 16:15:20 +02:00
// it mod basepoint to make it happy:
sc_reduce32 ( pk_sh_data ) ;
std : : memcpy ( keys . key . data , pk_sh_data , 32 ) ;
2020-09-02 16:39:25 +02:00
if ( ! crypto : : secret_key_to_public_key ( keys . key , keys . pub ) )
2020-09-03 02:38:03 +02:00
throw std : : runtime_error { " Failed to derive primary key from ed25519 key " } ;
2020-09-23 02:20:33 +02:00
assert ( 0 = = std : : memcmp ( keys . pub . data , keys . pub_ed25519 . data , 32 ) ) ;
2020-10-22 19:55:33 +02:00
} else if ( ! init_key ( m_config_folder / " key " , keys . key , keys . pub ,
2020-08-29 21:56:19 +02:00
crypto : : secret_key_to_public_key ,
[ ] ( crypto : : secret_key & key , crypto : : public_key & pubkey ) {
throw std : : runtime_error { " Internal error: old-style public keys are no longer generated " } ;
} ) )
return false ;
} else {
keys . key = crypto : : null_skey ;
keys . pub = crypto : : null_pkey ;
}
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( m_service_node ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " Service node public keys: " ) ) ;
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " - primary: {} " , tools : : type_to_hex ( keys . pub ) ) ) ;
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " - ed25519: {} " , tools : : type_to_hex ( keys . pub_ed25519 ) ) ) ;
C++17
Switch loki dev branch to C++17 compilation, and update the code with
various C++17 niceties.
- stop including the (deprecated) lokimq/string_view.h header and
instead switch everything to use std::string_view and `""sv` instead of
`""_sv`.
- std::string_view is much nicer than epee::span, so updated various
loki-specific code to use it instead.
- made epee "portable storage" serialization accept a std::string_view
instead of const lvalue std::string so that we can avoid copying.
- switched from mapbox::variant to std::variant
- use `auto [a, b] = whatever()` instead of `T1 a; T2 b; std::tie(a, b)
= whatever()` in a couple places (in the wallet code).
- switch to std::lock(...) instead of boost::lock(...) for simultaneous
lock acquisition. boost::lock() won't compile in C++17 mode when given
locks of different types.
- removed various pre-C++17 workarounds, e.g. for fold expressions,
unused argument attributes, and byte-spannable object detection.
- class template deduction means lock types no longer have to specify
the mutex, so `std::unique_lock<std::mutex> lock{mutex}` can become
`std::unique_lock lock{mutex}`. This will make switching any mutex
types (e.g. from boost to std mutexes) far easier as you just have to
update the type in the header and everything should work. This also
makes the tools::unique_lock and tools::shared_lock methods redundant
(which were a sort of poor-mans-pre-C++17 way to eliminate the
redundancy) so they are now gone and replaced with direct unique_lock or
shared_lock constructions.
- Redid the LNS validation using a string_view; instead of using raw
char pointers the code now uses a string view and chops off parts of the
view as it validates. So, for instance, it starts with "abcd.loki",
validates the ".loki" and chops the view to "abcd", then validates the
first character and chops to "bcd", validates the last and chops to
"bc", then can just check everything remaining for is-valid-middle-char.
- LNS validation gained a couple minor validation checks in the process:
- slightly tightened the requirement on lokinet addresses to require
that the last character of the mapped address is 'y' or 'o' (the
last base32z char holds only one significant bit).
- In parse_owner_to_generic_owner made sure that the owner value has
the correct size (otherwise we could up end not filling or
overfilling the pubkey buffer).
- Replaced base32z/base64/hex conversions with lokimq's versions which
have a nicer interface, are better optimized, and don't depend on epee.
2020-05-13 20:12:49 +02:00
// .snode address is the ed25519 pubkey, encoded with base32z and with .snode appended:
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " - lokinet: {}.snode " , oxenc : : to_base32z ( tools : : view_guts ( keys . pub_ed25519 ) ) ) ) ;
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " - x25519: {} " , tools : : type_to_hex ( keys . pub_x25519 ) ) ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
} else {
2020-05-11 22:00:58 +02:00
// Only print the x25519 version because it's the only thing useful for a non-SN (for
// encrypted LMQ RPC connections).
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " x25519 public key: {} " , tools : : type_to_hex ( keys . pub_x25519 ) ) ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
2018-07-12 04:07:34 +02:00
return true ;
}
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
2021-04-08 18:38:51 +02:00
oxenmq : : AuthLevel core : : omq_check_access ( const crypto : : x25519_public_key & pubkey ) const {
auto it = m_omq_auth . find ( pubkey ) ;
if ( it ! = m_omq_auth . end ( ) )
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
return it - > second ;
2021-01-14 21:28:50 +01:00
return oxenmq : : AuthLevel : : denied ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
// Builds an allow function; takes `*this`, the default auth level, and whether this connection
// should allow incoming SN connections.
//
// default_auth should be AuthLevel::denied if only pre-approved connections may connect,
// AuthLevel::basic for public RPC, AuthLevel::admin for a (presumably localhost) unrestricted
// port, and AuthLevel::none for a super restricted mode (generally this is useful when there are
// also SN-restrictions on commands, i.e. for quorumnet).
//
// check_sn is whether we check an incoming key against known service nodes (and thus return
// "true" for the service node access if it checks out).
//
2021-04-08 18:38:51 +02:00
oxenmq : : AuthLevel core : : omq_allow ( std : : string_view ip , std : : string_view x25519_pubkey_str , oxenmq : : AuthLevel default_auth ) {
2021-01-14 21:28:50 +01:00
using namespace oxenmq ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
AuthLevel auth = default_auth ;
if ( x25519_pubkey_str . size ( ) = = sizeof ( crypto : : x25519_public_key ) ) {
crypto : : x25519_public_key x25519_pubkey ;
std : : memcpy ( x25519_pubkey . data , x25519_pubkey_str . data ( ) , x25519_pubkey_str . size ( ) ) ;
2021-04-08 18:38:51 +02:00
auto user_auth = omq_check_access ( x25519_pubkey ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( user_auth > = AuthLevel : : basic ) {
if ( user_auth > auth )
auth = user_auth ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( oxen : : log : : Cat ( " omq " ) , " Incoming {}-authenticated connection " , auth ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( oxen : : log : : Cat ( " omq " ) , " Incoming [{}] curve connection from {}/{} " , auth , ip , x25519_pubkey ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
else {
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( oxen : : log : : Cat ( " omq " ) , " Incoming [{}] plain connection from {} " , auth , ip ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
return auth ;
}
2021-01-14 21:28:50 +01:00
void core : : init_oxenmq ( const boost : : program_options : : variables_map & vm ) {
using namespace oxenmq ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( omqlogcat , " Starting oxenmq " ) ;
2021-04-08 18:38:51 +02:00
m_omq = std : : make_unique < OxenMQ > (
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
tools : : copy_guts ( m_service_keys . pub_x25519 ) ,
tools : : copy_guts ( m_service_keys . key_x25519 ) ,
m_service_node ,
C++17
Switch loki dev branch to C++17 compilation, and update the code with
various C++17 niceties.
- stop including the (deprecated) lokimq/string_view.h header and
instead switch everything to use std::string_view and `""sv` instead of
`""_sv`.
- std::string_view is much nicer than epee::span, so updated various
loki-specific code to use it instead.
- made epee "portable storage" serialization accept a std::string_view
instead of const lvalue std::string so that we can avoid copying.
- switched from mapbox::variant to std::variant
- use `auto [a, b] = whatever()` instead of `T1 a; T2 b; std::tie(a, b)
= whatever()` in a couple places (in the wallet code).
- switch to std::lock(...) instead of boost::lock(...) for simultaneous
lock acquisition. boost::lock() won't compile in C++17 mode when given
locks of different types.
- removed various pre-C++17 workarounds, e.g. for fold expressions,
unused argument attributes, and byte-spannable object detection.
- class template deduction means lock types no longer have to specify
the mutex, so `std::unique_lock<std::mutex> lock{mutex}` can become
`std::unique_lock lock{mutex}`. This will make switching any mutex
types (e.g. from boost to std mutexes) far easier as you just have to
update the type in the header and everything should work. This also
makes the tools::unique_lock and tools::shared_lock methods redundant
(which were a sort of poor-mans-pre-C++17 way to eliminate the
redundancy) so they are now gone and replaced with direct unique_lock or
shared_lock constructions.
- Redid the LNS validation using a string_view; instead of using raw
char pointers the code now uses a string view and chops off parts of the
view as it validates. So, for instance, it starts with "abcd.loki",
validates the ".loki" and chops the view to "abcd", then validates the
first character and chops to "bcd", validates the last and chops to
"bc", then can just check everything remaining for is-valid-middle-char.
- LNS validation gained a couple minor validation checks in the process:
- slightly tightened the requirement on lokinet addresses to require
that the last character of the mapped address is 'y' or 'o' (the
last base32z char holds only one significant bit).
- In parse_owner_to_generic_owner made sure that the owner value has
the correct size (otherwise we could up end not filling or
overfilling the pubkey buffer).
- Replaced base32z/base64/hex conversions with lokimq's versions which
have a nicer interface, are better optimized, and don't depend on epee.
2020-05-13 20:12:49 +02:00
[ this ] ( std : : string_view x25519_pk ) { return m_service_node_list . remote_lookup ( x25519_pk ) ; } ,
2022-09-13 08:57:56 +02:00
[ ] ( LogLevel omqlevel , const char * file , int line , std : : string msg ) {
auto level = * oxen : : logging : : parse_level ( omqlevel ) ;
if ( omqlogcat - > should_log ( level ) )
omqlogcat - > log ( { file , line , " omq " } , level , " {} " , msg ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
} ,
2021-01-14 21:28:50 +01:00
oxenmq : : LogLevel : : trace
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
) ;
2021-04-08 18:38:51 +02:00
// ping.ping: a simple debugging target for pinging the omq listener
m_omq - > add_category ( " ping " , Access { AuthLevel : : none } )
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
. add_request_command ( " ping " , [ ] ( Message & m ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( oxen : : log : : Cat ( " omq " ) , " Received ping from {} " , m . conn ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
m . send_reply ( " pong " ) ;
} )
;
if ( m_service_node )
{
// Service nodes always listen for quorumnet data on the p2p IP, quorumnet port
std : : string listen_ip = vm [ " p2p-bind-ip " ] . as < std : : string > ( ) ;
if ( listen_ip . empty ( ) )
listen_ip = " 0.0.0.0 " ;
std : : string qnet_listen = " tcp:// " + listen_ip + " : " + std : : to_string ( m_quorumnet_port ) ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " - listening on {} (quorumnet) " , qnet_listen ) ;
2021-04-08 18:38:51 +02:00
m_omq - > listen_curve ( qnet_listen ,
[ this , public_ = command_line : : get_arg ( vm , arg_omq_quorumnet_public ) ] ( std : : string_view ip , std : : string_view pk , bool ) {
return omq_allow ( ip , pk , public_ ? AuthLevel : : basic : AuthLevel : : none ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
} ) ;
m_quorumnet_state = quorumnet_new ( * this ) ;
}
2020-08-21 09:01:37 +02:00
quorumnet_init ( * this , m_quorumnet_state ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
}
2021-01-14 21:28:50 +01:00
void core : : start_oxenmq ( ) {
2021-04-08 18:38:51 +02:00
update_omq_sns ( ) ; // Ensure we have SNs set for the current block before starting
2020-07-27 07:54:02 +02:00
if ( m_service_node )
{
2021-04-08 18:38:51 +02:00
m_pulse_thread_id = m_omq - > add_tagged_thread ( " pulse " ) ;
m_omq - > add_timer ( [ this ] ( ) { pulse : : main ( m_quorumnet_state , * this ) ; } ,
2020-08-10 11:37:17 +02:00
std : : chrono : : milliseconds ( 500 ) ,
false ,
m_pulse_thread_id ) ;
2021-04-08 18:38:51 +02:00
m_omq - > add_timer ( [ this ] ( ) { this - > check_service_node_time ( ) ; } ,
2020-12-01 01:56:43 +01:00
5 s ,
false ) ;
2020-07-27 07:54:02 +02:00
}
2021-04-08 18:38:51 +02:00
m_omq - > start ( ) ;
2020-05-11 23:37:24 +02:00
}
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
2018-07-12 04:07:34 +02:00
//-----------------------------------------------------------------------------------------------
2014-03-03 23:07:58 +01:00
bool core : : set_genesis_block ( const block & b )
{
return m_blockchain_storage . reset_and_set_genesis_block ( b ) ;
}
//-----------------------------------------------------------------------------------------------
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
void core : : deinit ( )
2014-03-03 23:07:58 +01:00
{
2020-01-19 23:44:18 +01:00
# ifdef ENABLE_SYSTEMD
sd_notify ( 0 , " STOPPING=1 \n STATUS=Shutting down " ) ;
# endif
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( m_quorumnet_state )
quorumnet_delete ( m_quorumnet_state ) ;
2021-04-08 18:38:51 +02:00
m_omq . reset ( ) ;
2019-01-18 06:16:23 +01:00
m_service_node_list . store ( ) ;
2015-12-14 05:54:39 +01:00
m_miner . stop ( ) ;
m_mempool . deinit ( ) ;
2016-10-03 03:06:55 +02:00
m_blockchain_storage . deinit ( ) ;
2014-03-03 23:07:58 +01:00
}
2015-02-12 20:59:39 +01:00
//-----------------------------------------------------------------------------------------------
2015-02-20 22:28:03 +01:00
void core : : test_drop_download ( )
2015-02-12 20:59:39 +01:00
{
2015-12-14 05:54:39 +01:00
m_test_drop_download = false ;
2015-02-12 20:59:39 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-02-20 22:28:03 +01:00
void core : : test_drop_download_height ( uint64_t height )
2015-02-12 20:59:39 +01:00
{
2015-12-14 05:54:39 +01:00
m_test_drop_download_height = height ;
2015-02-20 22:28:03 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : get_test_drop_download ( ) const
2015-02-20 22:28:03 +01:00
{
2015-12-14 05:54:39 +01:00
return m_test_drop_download ;
2015-02-20 22:28:03 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : get_test_drop_download_height ( ) const
2015-02-20 22:28:03 +01:00
{
2015-12-14 05:54:39 +01:00
if ( m_test_drop_download_height = = 0 )
return true ;
if ( get_blockchain_storage ( ) . get_current_blockchain_height ( ) < = m_test_drop_download_height )
return true ;
2015-02-20 22:28:03 +01:00
2015-12-14 05:54:39 +01:00
return false ;
2015-02-12 20:59:39 +01:00
}
2014-03-03 23:07:58 +01:00
//-----------------------------------------------------------------------------------------------
2019-11-25 05:57:15 +01:00
void core : : parse_incoming_tx_pre ( tx_verification_batch_info & tx_info )
2014-03-03 23:07:58 +01:00
{
2022-05-16 22:55:05 +02:00
if ( tx_info . blob - > size ( ) > MAX_TX_SIZE )
2014-03-03 23:07:58 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " WRONG TRANSACTION BLOB, too big size {}, rejected " , tx_info . blob - > size ( ) ) ;
2019-11-25 05:57:15 +01:00
tx_info . tvc . m_verifivation_failed = true ;
tx_info . tvc . m_too_big = true ;
return ;
2014-03-03 23:07:58 +01:00
}
2022-06-01 20:50:19 +02:00
else if ( tx_info . blob - > empty ( ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " WRONG TRANSACTION BLOB, blob is empty, rejected " ) ;
2022-06-01 20:50:19 +02:00
tx_info . tvc . m_verifivation_failed = true ;
return ;
}
2014-03-03 23:07:58 +01:00
2020-05-31 19:37:05 +02:00
tx_info . parsed = parse_and_validate_tx_from_blob ( * tx_info . blob , tx_info . tx , tx_info . tx_hash ) ;
2019-11-25 05:57:15 +01:00
if ( ! tx_info . parsed )
2014-03-03 23:07:58 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " WRONG TRANSACTION BLOB, Failed to parse, rejected " ) ;
2019-11-25 05:57:15 +01:00
tx_info . tvc . m_verifivation_failed = true ;
return ;
2014-03-03 23:07:58 +01:00
}
//std::cout << "!"<< tx.vin.size() << std::endl;
2020-06-02 01:01:24 +02:00
std : : lock_guard lock { bad_semantics_txes_lock } ;
2017-02-23 21:15:52 +01:00
for ( int idx = 0 ; idx < 2 ; + + idx )
2017-01-22 17:13:31 +01:00
{
2019-11-25 05:57:15 +01:00
if ( bad_semantics_txes [ idx ] . find ( tx_info . tx_hash ) ! = bad_semantics_txes [ idx ] . end ( ) )
2017-02-23 21:15:52 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Transaction already seen with bad semantics, rejected " ) ;
2019-11-25 05:57:15 +01:00
tx_info . tvc . m_verifivation_failed = true ;
return ;
2017-02-23 21:15:52 +01:00
}
2017-01-22 17:13:31 +01:00
}
2019-11-25 05:57:15 +01:00
tx_info . result = true ;
2017-07-10 16:38:56 +02:00
}
//-----------------------------------------------------------------------------------------------
2018-03-30 21:29:42 +02:00
void core : : set_semantics_failed ( const crypto : : hash & tx_hash )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " WRONG TRANSACTION BLOB, Failed to check tx {} semantic, rejected " , tx_hash ) ;
2018-03-30 21:29:42 +02:00
bad_semantics_txes_lock . lock ( ) ;
bad_semantics_txes [ 0 ] . insert ( tx_hash ) ;
if ( bad_semantics_txes [ 0 ] . size ( ) > = BAD_SEMANTICS_TXES_MAX_SIZE )
{
std : : swap ( bad_semantics_txes [ 0 ] , bad_semantics_txes [ 1 ] ) ;
bad_semantics_txes [ 0 ] . clear ( ) ;
}
bad_semantics_txes_lock . unlock ( ) ;
}
//-----------------------------------------------------------------------------------------------
2018-03-31 11:49:53 +02:00
static bool is_canonical_bulletproof_layout ( const std : : vector < rct : : Bulletproof > & proofs )
{
2018-07-18 23:24:53 +02:00
if ( proofs . size ( ) ! = 1 )
return false ;
const size_t sz = proofs [ 0 ] . V . size ( ) ;
2022-05-16 22:55:05 +02:00
if ( sz = = 0 | | sz > TX_BULLETPROOF_MAX_OUTPUTS )
2018-07-18 23:24:53 +02:00
return false ;
2018-03-31 11:49:53 +02:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2019-11-25 05:57:15 +01:00
void core : : parse_incoming_tx_accumulated_batch ( std : : vector < tx_verification_batch_info > & tx_info , bool kept_by_block )
2018-03-30 21:29:42 +02:00
{
2019-11-25 05:57:15 +01:00
if ( kept_by_block & & get_blockchain_storage ( ) . is_within_compiled_block_hash_area ( ) )
2017-07-30 10:45:55 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : trace ( logcat , " Skipping semantics check for txs kept by block in embedded hash area " ) ;
2019-11-25 05:57:15 +01:00
return ;
2017-07-30 10:45:55 +02:00
}
2018-03-30 21:29:42 +02:00
std : : vector < const rct : : rctSig * > rvv ;
for ( size_t n = 0 ; n < tx_info . size ( ) ; + + n )
2014-03-03 23:07:58 +01:00
{
2019-11-25 05:57:15 +01:00
if ( ! tx_info [ n ] . result | | tx_info [ n ] . already_have )
continue ;
if ( ! check_tx_semantic ( tx_info [ n ] . tx , kept_by_block ) )
2017-02-23 21:15:52 +01:00
{
2018-03-30 21:29:42 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
continue ;
}
2019-11-29 04:36:43 +01:00
if ( ! tx_info [ n ] . tx . is_transfer ( ) )
2018-03-30 21:29:42 +02:00
continue ;
2019-11-25 05:57:15 +01:00
const rct : : rctSig & rv = tx_info [ n ] . tx . rct_signatures ;
2018-03-30 21:29:42 +02:00
switch ( rv . type ) {
2020-11-21 05:20:15 +01:00
case rct : : RCTType : : Null :
2018-03-30 21:29:42 +02:00
// coinbase should not come here, so we reject for all other types
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Unexpected Null rctSig type " ) ;
2018-03-30 21:29:42 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
break ;
2020-11-21 05:20:15 +01:00
case rct : : RCTType : : Simple :
2018-03-30 21:29:42 +02:00
if ( ! rct : : verRctSemanticsSimple ( rv ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " rct signature semantics check failed " ) ;
2018-03-30 21:29:42 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
break ;
}
break ;
2020-11-21 05:20:15 +01:00
case rct : : RCTType : : Full :
2018-03-30 21:29:42 +02:00
if ( ! rct : : verRct ( rv , true ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " rct signature semantics check failed " ) ;
2018-03-30 21:29:42 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
break ;
}
break ;
2020-11-21 05:20:15 +01:00
case rct : : RCTType : : Bulletproof :
case rct : : RCTType : : Bulletproof2 :
case rct : : RCTType : : CLSAG :
2018-03-31 11:49:53 +02:00
if ( ! is_canonical_bulletproof_layout ( rv . p . bulletproofs ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Bulletproof does not have canonical form " ) ;
2018-03-31 11:49:53 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
break ;
}
2018-03-30 21:29:42 +02:00
rvv . push_back ( & rv ) ; // delayed batch verification
break ;
default :
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Unknown rct type: {} " , ( int ) rv . type ) ;
2018-03-30 21:29:42 +02:00
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
break ;
}
}
if ( ! rvv . empty ( ) & & ! rct : : verRctSemanticsSimple ( rvv ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " One transaction among this group has bad semantics, verifying one at a time " ) ;
2018-03-30 21:29:42 +02:00
const bool assumed_bad = rvv . size ( ) = = 1 ; // if there's only one tx, it must be the bad one
for ( size_t n = 0 ; n < tx_info . size ( ) ; + + n )
{
2019-11-25 05:57:15 +01:00
if ( ! tx_info [ n ] . result | | tx_info [ n ] . already_have )
2018-03-30 21:29:42 +02:00
continue ;
2020-09-17 01:21:48 +02:00
if ( ! rct : : is_rct_bulletproof ( tx_info [ n ] . tx . rct_signatures . type ) )
2018-03-30 21:29:42 +02:00
continue ;
2019-11-25 05:57:15 +01:00
if ( assumed_bad | | ! rct : : verRctSemanticsSimple ( tx_info [ n ] . tx . rct_signatures ) )
2018-03-30 21:29:42 +02:00
{
set_semantics_failed ( tx_info [ n ] . tx_hash ) ;
tx_info [ n ] . tvc . m_verifivation_failed = true ;
tx_info [ n ] . result = false ;
}
2017-02-23 21:15:52 +01:00
}
2014-03-03 23:07:58 +01:00
}
2017-07-10 16:38:56 +02:00
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
std : : vector < cryptonote : : tx_verification_batch_info > core : : parse_incoming_txs ( const std : : vector < std : : string > & tx_blobs , const tx_pool_options & opts )
2017-07-10 16:38:56 +02:00
{
2019-11-25 19:22:30 +01:00
// Caller needs to do this around both this *and* handle_parsed_txs
//auto lock = incoming_tx_lock();
2022-04-29 01:51:14 +02:00
std : : vector < cryptonote : : tx_verification_batch_info > tx_info ( tx_blobs . size ( ) ) ;
2017-07-10 16:38:56 +02:00
2018-04-23 19:55:58 +02:00
tools : : threadpool & tpool = tools : : threadpool : : getInstance ( ) ;
2017-09-14 05:39:37 +02:00
tools : : threadpool : : waiter waiter ;
2019-11-25 05:57:15 +01:00
for ( size_t i = 0 ; i < tx_blobs . size ( ) ; i + + ) {
tx_info [ i ] . blob = & tx_blobs [ i ] ;
tpool . submit ( & waiter , [ this , & info = tx_info [ i ] ] {
2017-09-14 05:39:37 +02:00
try
{
2019-11-25 05:57:15 +01:00
parse_incoming_tx_pre ( info ) ;
2017-09-14 05:39:37 +02:00
}
catch ( const std : : exception & e )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Exception in handle_incoming_tx_pre: {} " , e . what ( ) ) ;
2019-11-25 05:57:15 +01:00
info . tvc . m_verifivation_failed = true ;
2017-09-14 05:39:37 +02:00
}
} ) ;
}
2018-04-26 12:44:47 +02:00
waiter . wait ( & tpool ) ;
2019-10-23 06:13:55 +02:00
2019-11-25 05:57:15 +01:00
for ( auto & info : tx_info ) {
if ( ! info . result )
2017-09-14 05:39:37 +02:00
continue ;
2019-10-23 06:13:55 +02:00
2019-11-25 05:57:15 +01:00
if ( m_mempool . have_tx ( info . tx_hash ) )
2017-09-14 05:39:37 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " tx {} already has a transaction in tx_pool " , info . tx_hash ) ;
2019-11-25 05:57:15 +01:00
info . already_have = true ;
2017-09-14 05:39:37 +02:00
}
2019-11-25 05:57:15 +01:00
else if ( m_blockchain_storage . have_tx ( info . tx_hash ) )
2017-09-14 05:39:37 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " tx {} already has a transaction in tx_pool " , info . tx_hash ) ;
2019-11-25 05:57:15 +01:00
info . already_have = true ;
2017-07-10 16:38:56 +02:00
}
2018-03-30 21:29:42 +02:00
}
2019-10-23 06:13:55 +02:00
2021-02-08 06:31:05 +01:00
2019-11-25 05:57:15 +01:00
parse_incoming_tx_accumulated_batch ( tx_info , opts . kept_by_block ) ;
2018-03-30 21:29:42 +02:00
2019-11-25 05:57:15 +01:00
return tx_info ;
}
2018-03-30 21:29:42 +02:00
2019-11-25 19:22:30 +01:00
bool core : : handle_parsed_txs ( std : : vector < tx_verification_batch_info > & parsed_txs , const tx_pool_options & opts ,
2019-11-25 05:57:15 +01:00
uint64_t * blink_rollback_height )
{
2019-11-25 19:22:30 +01:00
// Caller needs to do this around both this *and* parse_incoming_txs
//auto lock = incoming_tx_lock();
2022-05-16 22:55:05 +02:00
auto version = m_blockchain_storage . get_network_version ( ) ;
bool ok = true ;
2019-11-25 05:57:15 +01:00
if ( blink_rollback_height )
* blink_rollback_height = 0 ;
tx_pool_options tx_opts ;
for ( size_t i = 0 ; i < parsed_txs . size ( ) ; i + + ) {
auto & info = parsed_txs [ i ] ;
if ( ! info . result )
2017-07-10 16:38:56 +02:00
{
2019-11-25 05:57:15 +01:00
ok = false ; // Propagate failures (so this can be chained with parse_incoming_txs without an intermediate check)
2017-07-10 16:38:56 +02:00
continue ;
}
Generic burn fee checking + blink burn fee checking
This adds the ability for check_fee() to also check the burn amount.
This requires passing extra info through `add_tx()` (and the various
things that call it), so I took the:
bool keeped_by_block, bool relayed, bool do_not_relay
argument triplet, moved it into a struct in tx_pool.h, then added the other fee
options there (along with some static factory functions for generating the
typical sets of option).
The majority of this commit is chasing that change through the codebase and
test suite.
This is used by blink but should also help LNS and other future burn
transactions to verify a burn amount simply when adding the transation to the
mempool. It supports a fixed burn amount, a burn amount as a multiple of the
minimum tx fee, and also allows you to increase the minimum tx fee (so that,
for example, we could require blink txes to pay miners 250% of the usual
minimum (unimportant) priority tx fee.
- Removed a useless core::add_new_tx() overload that wasn't used anywhere.
Blink-specific changes:
(I'd normally separate these into a separate commit, but they got interwoven
fairly heavily with the above change).
- changed the way blink burning is specified so that we have three knobs for
fee adjustment (fixed burn fee; base fee multiple; and required miner tx fee).
The fixed amount is currently 0, base fee is 400%, and require miner tx fee is
simply 100% (i.e. no different than a normal transaction). This is the same as
before this commit, but is changing how they are being specified in
cryptonote_config.h.
- blink tx fee, burn amount, and miner tx fee (if > 100%) now get checked
before signing a blink tx. (These fee checks don't apply to anyone else --
when propagating over the network only the miner tx fee is checked).
- Added a couple of checks for blink quorums: 1) make sure they have reached
the blink hf; 2) make sure the submitted tx version conforms to the current hf
min/max tx version.
- print blink fee information in simplewallet's `fee` output
- add "typical" fee calculations in the `fee` output:
[wallet T6SCwL (has locked stakes)]: fee
Current fee is 0.000000850 loki per byte + 0.020000000 loki per output
No backlog at priority 1
No backlog at priority 2
No backlog at priority 3
No backlog at priority 4
Current blink fee is 0.000004250 loki per byte + 0.100000000 loki per output
Estimated typical small transaction fees: 0.042125000 (unimportant), 0.210625000 (normal), 1.053125000 (elevated), 5.265625000 (priority), 0.210625000 (blink)
where "small" here is the same tx size (2500 bytes + 2 outputs) used to
estimate backlogs.
2019-11-09 04:14:15 +01:00
if ( opts . kept_by_block )
2019-11-25 05:57:15 +01:00
get_blockchain_storage ( ) . on_new_tx_from_block ( info . tx ) ;
if ( info . already_have )
continue ; // Not a failure
2017-07-10 16:38:56 +02:00
2019-11-25 05:57:15 +01:00
const size_t weight = get_transaction_weight ( info . tx , info . blob - > size ( ) ) ;
const tx_pool_options * local_opts = & opts ;
if ( blink_rollback_height & & info . approved_blink )
2018-11-21 06:30:16 +01:00
{
2019-11-25 05:57:15 +01:00
// If this is an approved blink then pass a copy of the options with the flag added
tx_opts = opts ;
tx_opts . approved_blink = true ;
local_opts = & tx_opts ;
2018-11-21 06:30:16 +01:00
}
2019-11-25 05:57:15 +01:00
if ( m_mempool . add_tx ( info . tx , info . tx_hash , * info . blob , weight , info . tvc , * local_opts , version , blink_rollback_height ) )
2019-12-18 02:49:56 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " tx added: {} " , info . tx_hash ) ;
2019-12-18 02:49:56 +01:00
}
2019-11-25 05:57:15 +01:00
else
2018-11-21 06:30:16 +01:00
{
2019-11-25 05:57:15 +01:00
ok = false ;
if ( info . tvc . m_verifivation_failed )
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Transaction verification failed: {} " , info . tx_hash ) ;
2019-11-25 05:57:15 +01:00
else if ( info . tvc . m_verifivation_impossible )
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " Transaction verification impossible: {} " , info . tx_hash ) ;
2018-11-21 06:30:16 +01:00
}
2017-07-10 16:38:56 +02:00
}
2017-08-29 14:01:45 +02:00
2019-11-25 05:57:15 +01:00
return ok ;
2017-07-10 16:38:56 +02:00
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
std : : vector < cryptonote : : tx_verification_batch_info > core : : handle_incoming_txs ( const std : : vector < std : : string > & tx_blobs , const tx_pool_options & opts )
2019-11-25 19:22:30 +01:00
{
auto lock = incoming_tx_lock ( ) ;
auto parsed = parse_incoming_txs ( tx_blobs , opts ) ;
handle_parsed_txs ( parsed , opts ) ;
return parsed ;
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : handle_incoming_tx ( const std : : string & tx_blob , tx_verification_context & tvc , const tx_pool_options & opts )
2017-07-10 16:38:56 +02:00
{
2022-05-20 23:29:48 +02:00
const std : : vector < std : : string > tx_blobs { { tx_blob } } ;
2019-11-25 19:22:30 +01:00
auto parsed = handle_incoming_txs ( tx_blobs , opts ) ;
parsed [ 0 ] . blob = & tx_blob ; // Update pointer to the input rather than the copy in case the caller wants to use it for some reason
2019-11-25 05:57:15 +01:00
tvc = parsed [ 0 ] . tvc ;
2019-11-25 19:22:30 +01:00
return parsed [ 0 ] . result & & ( parsed [ 0 ] . already_have | | tvc . m_added_to_pool ) ;
2014-03-03 23:07:58 +01:00
}
2019-11-06 07:28:33 +01:00
//-----------------------------------------------------------------------------------------------
2019-11-25 05:57:15 +01:00
std : : pair < std : : vector < std : : shared_ptr < blink_tx > > , std : : unordered_set < crypto : : hash > >
core : : parse_incoming_blinks ( const std : : vector < serializable_blink_metadata > & blinks )
2019-11-06 07:28:33 +01:00
{
2019-11-25 05:57:15 +01:00
std : : pair < std : : vector < std : : shared_ptr < blink_tx > > , std : : unordered_set < crypto : : hash > > results ;
auto & new_blinks = results . first ;
auto & missing_txs = results . second ;
2019-11-18 23:38:16 +01:00
2022-05-16 22:55:05 +02:00
if ( m_blockchain_storage . get_network_version ( ) < feature : : BLINK )
2019-11-25 05:57:15 +01:00
return results ;
2017-07-10 16:38:56 +02:00
2019-11-25 05:57:15 +01:00
std : : vector < uint8_t > want ( blinks . size ( ) , false ) ; // Really bools, but std::vector<bool> is broken.
size_t want_count = 0 ;
// Step 1: figure out which referenced transactions we want to keep:
// - unknown tx (typically an incoming blink)
// - in mempool without blink sigs (it's possible to get the tx before the blink signatures)
// - in a recent, still-mutable block with blink sigs (can happen when syncing blocks before
// retrieving blink signatures)
2019-11-06 07:28:33 +01:00
{
std : : vector < crypto : : hash > hashes ;
hashes . reserve ( blinks . size ( ) ) ;
for ( auto & bm : blinks )
hashes . emplace_back ( bm . tx_hash ) ;
2019-11-26 16:53:16 +01:00
std : : unique_lock < Blockchain > lock ( m_blockchain_storage ) ;
2019-11-06 07:28:33 +01:00
auto tx_block_heights = m_blockchain_storage . get_transactions_heights ( hashes ) ;
auto immutable_height = m_blockchain_storage . get_immutable_height ( ) ;
auto & db = m_blockchain_storage . get_db ( ) ;
for ( size_t i = 0 ; i < blinks . size ( ) ; i + + ) {
2019-11-25 05:57:15 +01:00
if ( tx_block_heights [ i ] = = 0 /*mempool or unknown*/ | | tx_block_heights [ i ] > immutable_height /*mined but not yet immutable*/ )
{
want [ i ] = true ;
want_count + + ;
}
2019-11-06 07:28:33 +01:00
}
}
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " Want {} of {} incoming blink signature sets after filtering out immutable txes " , want_count , blinks . size ( ) ) ;
2019-11-25 05:57:15 +01:00
if ( ! want_count ) return results ;
2019-11-06 07:28:33 +01:00
// Step 2: filter out any transactions for which we already have a blink signature
{
auto mempool_lock = m_mempool . blink_shared_lock ( ) ;
for ( size_t i = 0 ; i < blinks . size ( ) ; i + + )
2018-11-21 06:30:16 +01:00
{
Fix rare blink pool deadlock
cryptonote_protocol_handler calls `pool.get_blink(hash)` while already
holding a blink shared lock, which should have been
`pool.get_blink(hash, true)` to avoid `get_blink` trying to take its own
lock.
That double lock is undefined behaviour and can cause a deadlock on the
mutex, although it appears rare that it actually does. If it does,
however, this eventually backs up into vote relaying during the idle
loop, which then stalls the idle loop so we stop sending out uptime
proofs (since that is also in the idle loop).
A simple fix here is to add the `true` argument, but on reconsideration
this extra argument to take or not take a lock is messy and error prone,
so this commit instead removes the second argument entirely and instead
documents which call must and must not hold a lock, getting rid of the
three methods (get_blink, has_blink, and add_existing_blink) that had
the `have_lock` argument. This ends up having only a small impact on
calling code - the vast majority of callers already hold a lock, and the
few that don't are easily adjusted.
2020-01-18 19:01:45 +01:00
if ( want [ i ] & & m_mempool . has_blink ( blinks [ i ] . tx_hash ) )
2019-11-06 07:28:33 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " Ignoring blink data for {}: already have blink signatures " , blinks [ i ] . tx_hash ) ;
2019-11-25 05:57:15 +01:00
want [ i ] = false ; // Already have it, move along
want_count - - ;
2019-11-06 07:28:33 +01:00
}
}
}
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " Want {} of {} incoming blink signature sets after filtering out existing blink sigs " , want_count , blinks . size ( ) ) ;
2019-11-25 05:57:15 +01:00
if ( ! want_count ) return results ;
// Step 3: create new blink_tx objects for txes and add the blink signatures. We can do all of
// this without a lock since these are (for now) just local instances.
new_blinks . reserve ( want_count ) ;
2019-11-06 07:28:33 +01:00
std : : unordered_map < uint64_t , std : : shared_ptr < const service_nodes : : quorum > > quorum_cache ;
for ( size_t i = 0 ; i < blinks . size ( ) ; i + + )
{
2019-11-25 05:57:15 +01:00
if ( ! want [ i ] )
2019-11-06 07:28:33 +01:00
continue ;
auto & bdata = blinks [ i ] ;
2019-11-25 05:57:15 +01:00
new_blinks . push_back ( std : : make_shared < blink_tx > ( bdata . height , bdata . tx_hash ) ) ;
auto & blink = * new_blinks . back ( ) ;
2019-11-20 01:24:48 +01:00
// Data structure checks (we have more stringent checks for validity later, but if these fail
2019-11-25 05:57:15 +01:00
// now then there's no point of even trying to do signature validation.
2019-11-20 01:24:48 +01:00
if ( bdata . signature . size ( ) ! = bdata . position . size ( ) | | // Each signature must have an associated quorum position
bdata . signature . size ( ) ! = bdata . quorum . size ( ) | | // and quorum index
bdata . signature . size ( ) < service_nodes : : BLINK_MIN_VOTES * tools : : enum_count < blink_tx : : subquorum > | | // too few signatures for possible validity
bdata . signature . size ( ) > service_nodes : : BLINK_SUBQUORUM_SIZE * tools : : enum_count < blink_tx : : subquorum > | | // too many signatures
blink_tx : : quorum_height ( bdata . height , blink_tx : : subquorum : : base ) = = 0 | | // Height is too early (no blink quorum height)
std : : any_of ( bdata . position . begin ( ) , bdata . position . end ( ) , [ ] ( const auto & p ) { return p > = service_nodes : : BLINK_SUBQUORUM_SIZE ; } ) | | // invalid position
std : : any_of ( bdata . quorum . begin ( ) , bdata . quorum . end ( ) , [ ] ( const auto & qi ) { return qi > = tools : : enum_count < blink_tx : : subquorum > ; } ) // invalid quorum index
2019-11-06 07:28:33 +01:00
) {
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Invalid blink tx {}: invalid signature data " , bdata . tx_hash ) ;
2019-11-06 07:28:33 +01:00
continue ;
2018-11-21 06:30:16 +01:00
}
2019-11-06 07:28:33 +01:00
2019-12-24 04:01:48 +01:00
bool no_quorum = false ;
2019-11-06 07:28:33 +01:00
std : : array < const std : : vector < crypto : : public_key > * , tools : : enum_count < blink_tx : : subquorum > > validators ;
for ( uint8_t qi = 0 ; qi < tools : : enum_count < blink_tx : : subquorum > ; qi + + )
2018-11-21 06:30:16 +01:00
{
2019-11-25 05:57:15 +01:00
auto q_height = blink . quorum_height ( static_cast < blink_tx : : subquorum > ( qi ) ) ;
2019-11-06 07:28:33 +01:00
auto & q = quorum_cache [ q_height ] ;
if ( ! q )
q = get_quorum ( service_nodes : : quorum_type : : blink , q_height ) ;
2019-12-24 04:01:48 +01:00
if ( ! q )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Don't have a quorum for height {} (yet?), ignoring this blink " , q_height ) ;
2019-12-24 04:01:48 +01:00
no_quorum = true ;
break ;
}
2019-11-06 07:28:33 +01:00
validators [ qi ] = & q - > validators ;
2018-11-21 06:30:16 +01:00
}
2019-12-24 04:01:48 +01:00
if ( no_quorum )
continue ;
2017-07-10 16:38:56 +02:00
2019-11-20 01:24:48 +01:00
std : : vector < std : : pair < size_t , std : : string > > failures ;
for ( size_t s = 0 ; s < bdata . signature . size ( ) ; s + + )
{
try {
2019-11-06 07:28:33 +01:00
blink . add_signature ( static_cast < blink_tx : : subquorum > ( bdata . quorum [ s ] ) , bdata . position [ s ] , true /*approved*/ , bdata . signature [ s ] ,
validators [ bdata . quorum [ s ] ] - > at ( bdata . position [ s ] ) ) ;
2019-11-20 01:24:48 +01:00
} catch ( const std : : exception & e ) {
failures . emplace_back ( s , e . what ( ) ) ;
}
}
if ( blink . approved ( ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Blink tx {} blink signatures approved with {} signature validation failures " , bdata . tx_hash , failures . size ( ) ) ;
2019-11-20 01:24:48 +01:00
for ( auto & f : failures )
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " - failure for quorum {}, position {}: {} " , int ( bdata . quorum [ f . first ] ) , int ( bdata . position [ f . first ] ) , f . second ) ;
2019-11-20 01:24:48 +01:00
}
else
{
std : : ostringstream os ;
os < < " Blink validation failed: " ;
for ( auto & f : failures )
os < < " [ " < < int ( bdata . quorum [ f . first ] ) < < " : " < < int ( bdata . position [ f . first ] ) < < " ]: " < < f . second ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Invalid blink tx {}: {} " , bdata . tx_hash , os . str ( ) ) ;
2019-11-06 07:28:33 +01:00
}
2017-07-10 16:38:56 +02:00
}
2017-08-29 14:01:45 +02:00
2019-11-25 05:57:15 +01:00
return results ;
2017-07-10 16:38:56 +02:00
}
2019-11-25 05:57:15 +01:00
int core : : add_blinks ( const std : : vector < std : : shared_ptr < blink_tx > > & blinks )
{
2019-11-20 01:24:48 +01:00
int added = 0 ;
2019-11-25 05:57:15 +01:00
if ( blinks . empty ( ) )
return added ;
auto lock = m_mempool . blink_unique_lock ( ) ;
for ( auto & b : blinks )
if ( b - > approved ( ) )
Fix rare blink pool deadlock
cryptonote_protocol_handler calls `pool.get_blink(hash)` while already
holding a blink shared lock, which should have been
`pool.get_blink(hash, true)` to avoid `get_blink` trying to take its own
lock.
That double lock is undefined behaviour and can cause a deadlock on the
mutex, although it appears rare that it actually does. If it does,
however, this eventually backs up into vote relaying during the idle
loop, which then stalls the idle loop so we stop sending out uptime
proofs (since that is also in the idle loop).
A simple fix here is to add the `true` argument, but on reconsideration
this extra argument to take or not take a lock is messy and error prone,
so this commit instead removes the second argument entirely and instead
documents which call must and must not hold a lock, getting rid of the
three methods (get_blink, has_blink, and add_existing_blink) that had
the `have_lock` argument. This ends up having only a small impact on
calling code - the vast majority of callers already hold a lock, and the
few that don't are easily adjusted.
2020-01-18 19:01:45 +01:00
if ( m_mempool . add_existing_blink ( b ) )
2019-11-20 01:24:48 +01:00
added + + ;
2019-11-06 07:28:33 +01:00
2020-10-01 10:08:19 +02:00
if ( added )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Added blink signatures for {} blinks " , added ) ;
2020-10-01 10:08:19 +02:00
long_poll_trigger ( m_mempool ) ;
}
2019-11-20 01:24:48 +01:00
2019-11-25 05:57:15 +01:00
return added ;
2019-11-06 07:28:33 +01:00
}
2017-07-10 16:38:56 +02:00
//-----------------------------------------------------------------------------------------------
2019-10-27 23:47:19 +01:00
std : : future < std : : pair < blink_result , std : : string > > core : : handle_blink_tx ( const std : : string & tx_blob )
2017-07-10 16:38:56 +02:00
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
return quorumnet_send_blink ( * this , tx_blob ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : check_tx_semantic ( const transaction & tx , bool keeped_by_block ) const
2014-03-03 23:07:58 +01:00
{
2019-11-29 04:36:43 +01:00
if ( tx . is_transfer ( ) )
2018-07-25 03:13:38 +02:00
{
2019-11-29 04:36:43 +01:00
if ( tx . vin . empty ( ) )
2018-07-25 03:13:38 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx with empty inputs, rejected for tx id= {} " , get_transaction_hash ( tx ) ) ;
2018-07-25 03:13:38 +02:00
return false ;
}
}
2019-11-29 04:36:43 +01:00
else
2014-03-03 23:07:58 +01:00
{
2019-11-29 04:36:43 +01:00
if ( tx . vin . size ( ) ! = 0 )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx type: {} must have 0 inputs, received: {}, rejected for tx id = {} " , tx . type , tx . vin . size ( ) , get_transaction_hash ( tx ) ) ;
2019-11-29 04:36:43 +01:00
return false ;
}
2014-03-03 23:07:58 +01:00
}
if ( ! check_inputs_types_supported ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " unsupported input types for tx id= {} " , get_transaction_hash ( tx ) ) ;
2014-03-03 23:07:58 +01:00
return false ;
}
if ( ! check_outs_valid ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx with invalid outputs, rejected for tx id= {} " , get_transaction_hash ( tx ) ) ;
2014-03-03 23:07:58 +01:00
return false ;
}
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
2019-06-11 20:53:46 +02:00
if ( tx . version > = txversion : : v2_ringct )
2016-06-16 00:37:13 +02:00
{
if ( tx . rct_signatures . outPk . size ( ) ! = tx . vout . size ( ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx with mismatched vout/outPk count, rejected for tx id= {} " , get_transaction_hash ( tx ) ) ;
2016-06-16 00:37:13 +02:00
return false ;
}
}
2014-03-03 23:07:58 +01:00
if ( ! check_money_overflow ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx has money overflow, rejected for tx id= {} " , get_transaction_hash ( tx ) ) ;
2014-03-03 23:07:58 +01:00
return false ;
}
2019-06-11 20:53:46 +02:00
if ( tx . version = = txversion : : v1 )
2014-03-03 23:07:58 +01:00
{
2016-06-16 00:37:13 +02:00
uint64_t amount_in = 0 ;
get_inputs_money_amount ( tx , amount_in ) ;
uint64_t amount_out = get_outs_money_amount ( tx ) ;
if ( amount_in < = amount_out )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx with wrong amounts: ins {}, outs {}, rejected for tx id= {} " , amount_in , amount_out , get_transaction_hash ( tx ) ) ;
2016-06-16 00:37:13 +02:00
return false ;
}
2014-03-03 23:07:58 +01:00
}
2022-05-16 22:55:05 +02:00
if ( ! keeped_by_block & & get_transaction_weight ( tx ) > = m_blockchain_storage . get_current_cumulative_block_weight_limit ( ) - COINBASE_BLOB_RESERVED_SIZE )
2014-03-03 23:07:58 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx is too large {}, expected not bigger than {} " , get_transaction_weight ( tx ) , m_blockchain_storage . get_current_cumulative_block_weight_limit ( ) - COINBASE_BLOB_RESERVED_SIZE ) ;
2014-03-03 23:07:58 +01:00
return false ;
}
if ( ! check_tx_inputs_keyimages_diff ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx uses a single key image more than once " ) ;
2014-03-03 23:07:58 +01:00
return false ;
}
2017-07-18 23:38:26 +02:00
if ( ! check_tx_inputs_ring_members_diff ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx uses duplicate ring members " ) ;
2017-07-18 23:38:26 +02:00
return false ;
}
2017-02-20 18:41:39 +01:00
if ( ! check_tx_inputs_keyimages_domain ( tx ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " tx uses key image not in the valid domain " ) ;
2017-02-20 18:41:39 +01:00
return false ;
}
2014-03-03 23:07:58 +01:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2020-12-01 01:56:43 +01:00
bool core : : check_service_node_time ( )
{
if ( ! is_active_sn ( ) ) { return true ; }
crypto : : public_key pubkey = m_service_node_list . get_random_pubkey ( ) ;
crypto : : x25519_public_key x_pkey { 0 } ;
2021-04-22 05:46:38 +02:00
constexpr std : : array < uint16_t , 3 > MIN_TIMESTAMP_VERSION { 9 , 1 , 0 } ;
2020-12-01 01:56:43 +01:00
std : : array < uint16_t , 3 > proofversion ;
2021-02-08 06:31:05 +01:00
m_service_node_list . access_proof ( pubkey , [ & ] ( auto & proof ) {
x_pkey = proof . pubkey_x25519 ;
2021-01-28 01:07:57 +01:00
proofversion = proof . proof - > version ;
2020-12-01 01:56:43 +01:00
} ) ;
if ( proofversion > = MIN_TIMESTAMP_VERSION & & x_pkey ) {
2021-04-08 18:38:51 +02:00
m_omq - > request (
2020-12-01 01:56:43 +01:00
tools : : view_guts ( x_pkey ) ,
" quorum.timestamp " ,
[ this , pubkey ] ( bool success , std : : vector < std : : string > data ) {
const time_t local_seconds = time ( nullptr ) ;
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " Timestamp message received: {}, local time is: " , data [ 0 ] , local_seconds ) ;
2021-02-08 06:31:05 +01:00
if ( success ) {
2020-12-01 01:56:43 +01:00
int64_t received_seconds ;
if ( tools : : parse_int ( data [ 0 ] , received_seconds ) ) {
uint16_t variance ;
if ( received_seconds > local_seconds + 65535 | | received_seconds < local_seconds - 65535 ) {
variance = 65535 ;
} else {
variance = std : : abs ( local_seconds - received_seconds ) ;
}
std : : lock_guard < std : : mutex > lk ( m_sn_timestamp_mutex ) ;
// Records the variance into the record of our performance (m_sn_times)
service_nodes : : timesync_entry entry { variance < = service_nodes : : THRESHOLD_SECONDS_OUT_OF_SYNC } ;
2021-02-08 06:31:05 +01:00
m_sn_times . add ( entry ) ;
2020-12-01 01:56:43 +01:00
// Counts the number of times we have been out of sync
2021-08-06 01:12:04 +02:00
if ( m_sn_times . failures ( ) > ( m_sn_times . size ( ) * service_nodes : : MAXIMUM_EXTERNAL_OUT_OF_SYNC / 100 ) ) {
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , " service node time might be out of sync " ) ;
2020-12-01 01:56:43 +01:00
// If we are out of sync record the other service node as in sync
m_service_node_list . record_timesync_status ( pubkey , true ) ;
} else {
m_service_node_list . record_timesync_status ( pubkey , variance < = service_nodes : : THRESHOLD_SECONDS_OUT_OF_SYNC ) ;
}
} else {
success = false ;
}
}
m_service_node_list . record_timestamp_participation ( pubkey , success ) ;
} ) ;
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : is_key_image_spent ( const crypto : : key_image & key_image ) const
2015-08-11 11:49:15 +02:00
{
return m_blockchain_storage . have_tx_keyimg_as_spent ( key_image ) ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : are_key_images_spent ( const std : : vector < crypto : : key_image > & key_im , std : : vector < bool > & spent ) const
2015-08-11 11:49:15 +02:00
{
spent . clear ( ) ;
2017-01-22 21:38:10 +01:00
for ( auto & ki : key_im )
2015-08-11 11:49:15 +02:00
{
spent . push_back ( m_blockchain_storage . have_tx_keyimg_as_spent ( ki ) ) ;
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2017-08-16 22:19:53 +02:00
size_t core : : get_block_sync_size ( uint64_t height ) const
{
2022-05-16 22:55:05 +02:00
return block_sync_size > 0 ? block_sync_size : BLOCKS_SYNCHRONIZING_DEFAULT_COUNT ;
2017-08-16 22:19:53 +02:00
}
//-----------------------------------------------------------------------------------------------
2017-09-05 18:20:27 +02:00
bool core : : are_key_images_spent_in_pool ( const std : : vector < crypto : : key_image > & key_im , std : : vector < bool > & spent ) const
{
spent . clear ( ) ;
return m_mempool . check_for_key_images ( key_im , spent ) ;
}
//-----------------------------------------------------------------------------------------------
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
std : : optional < std : : tuple < int64_t , int64_t , int64_t > > core : : get_coinbase_tx_sum ( uint64_t start_offset , size_t count )
2016-10-10 21:45:51 +02:00
{
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
std : : optional < std : : tuple < int64_t , int64_t , int64_t > > result { { 0 , 0 , 0 } } ;
2020-08-13 21:40:54 +02:00
if ( count = = 0 )
return result ;
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
auto & [ emission_amount , total_fee_amount , burnt_oxen ] = * result ;
2020-08-13 21:40:54 +02:00
// Caching.
//
// Requesting this value from the beginning of the chain is very slow, so we cache it. That
// still means the first request will be slow, but that's okay. To prevent a bunch of threads
// getting backed up trying to calculate this, we lock out more than one thread building the
// cache at a time if we're requesting a large number of block values at once. Any other thread
// requesting will get a nullopt back.
constexpr uint64_t CACHE_LAG = 30 ; // We cache the values up to this many blocks ago; we lag so that we don't have to worry about small reorgs
constexpr uint64_t CACHE_EXCLUSIVE = 1000 ; // If we need to load more than this, we block out other threads
// Check if we have a cacheable from-the-beginning result
uint64_t cache_to = 0 ;
std : : chrono : : steady_clock : : time_point cache_build_started ;
if ( start_offset = = 0 ) {
uint64_t height = m_blockchain_storage . get_current_blockchain_height ( ) ;
if ( count > height ) count = height ;
cache_to = height - std : : min ( CACHE_LAG , height ) ;
{
std : : shared_lock lock { m_coinbase_cache . mutex } ;
2021-02-19 17:23:10 +01:00
if ( m_coinbase_cache . height & & count > = m_coinbase_cache . height ) {
2020-08-13 21:40:54 +02:00
emission_amount = m_coinbase_cache . emissions ;
total_fee_amount = m_coinbase_cache . fees ;
2021-01-04 01:09:45 +01:00
burnt_oxen = m_coinbase_cache . burnt ;
2021-02-19 07:42:50 +01:00
start_offset = m_coinbase_cache . height + 1 ;
2020-08-13 21:40:54 +02:00
count - = m_coinbase_cache . height ;
}
// else don't change anything; we need a subset of blocks that ends before the cache.
if ( cache_to < = m_coinbase_cache . height )
cache_to = 0 ; // Cache doesn't need updating
}
// If we're loading a lot then acquire an exclusive lock, recheck our variables, and block out
// other threads until we're done. (We don't do this if we're only loading a few because even
// if we have some competing cache updates they don't hurt anything).
if ( cache_to > 0 & & count > CACHE_EXCLUSIVE ) {
std : : unique_lock lock { m_coinbase_cache . mutex } ;
if ( m_coinbase_cache . building )
return std : : nullopt ; // Another thread is already updating the cache
2021-02-19 17:23:10 +01:00
if ( m_coinbase_cache . height & & m_coinbase_cache . height > = start_offset ) {
2020-08-13 21:40:54 +02:00
// Someone else updated the cache while we were acquiring the unique lock, so update our variables
if ( m_coinbase_cache . height > = start_offset + count ) {
// The cache is now *beyond* us, which means we can't use it, so reset start/count back
// to what they were originally.
2021-02-19 07:42:50 +01:00
count + = start_offset - 1 ;
2020-08-13 21:40:54 +02:00
start_offset = 0 ;
cache_to = 0 ;
} else {
// The cache is updated and we can still use it, so update our variables.
emission_amount = m_coinbase_cache . emissions ;
total_fee_amount = m_coinbase_cache . fees ;
2021-01-04 01:09:45 +01:00
burnt_oxen = m_coinbase_cache . burnt ;
2021-02-19 07:42:50 +01:00
count - = m_coinbase_cache . height - start_offset + 1 ;
start_offset = m_coinbase_cache . height + 1 ;
2020-08-13 21:40:54 +02:00
}
}
if ( cache_to > 0 & & count > CACHE_EXCLUSIVE ) {
cache_build_started = std : : chrono : : steady_clock : : now ( ) ;
m_coinbase_cache . building = true ; // Block out other threads until we're done
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Starting slow cache build request for get_coinbase_tx_sum({}, {}) " , start_offset , count ) ;
2020-08-13 21:40:54 +02:00
}
}
}
const uint64_t end = start_offset + count - 1 ;
m_blockchain_storage . for_blocks_range ( start_offset , end ,
[ this , & cache_to , & result , & cache_build_started ] ( uint64_t height , const crypto : : hash & hash , const block & b ) {
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
auto & [ emission_amount , total_fee_amount , burnt_oxen ] = * result ;
2018-04-16 01:16:02 +02:00
std : : vector < transaction > txs ;
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
auto coinbase_amount = static_cast < int64_t > ( get_outs_money_amount ( b . miner_tx ) ) ;
2021-08-19 05:43:26 +02:00
get_transactions ( b . tx_hashes , txs ) ;
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
int64_t tx_fee_amount = 0 ;
2017-01-22 21:38:10 +01:00
for ( const auto & tx : txs )
2016-10-11 01:55:18 +02:00
{
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
tx_fee_amount + = static_cast < int64_t > ( get_tx_miner_fee ( tx , b . major_version > = feature : : FEE_BURNING ) ) ;
2022-05-16 22:55:05 +02:00
if ( b . major_version > = feature : : FEE_BURNING )
2020-03-12 10:53:24 +01:00
{
Fix negative coinbase emissions
With batching, individual blocks can have a negative coinbase emission
because the tx fee gets added to the batch rewards database and not paid
out immediately, which then results in an negative overflow to a value
close to 2^64. Thus a block with no payout and a tx fee will have an
erroneous huge positive coinbase emission when queried via
`get_coinbase_tx_sum`. For example block 1094068 queried with:
{"jsonrpc":"2.0","id":"0","method":"get_coinbase_tx_sum","params":{"height": 1094068, "count": 1}}
returns:
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": 18446744073699378616,
"fee_amount": 10173000,
"status": "OK"
}
}
This commit fixes it by making the values signed (and also serves as an
example of why unsigned integers are usually the wrong choice):
{
"jsonrpc": "2.0",
"id": "0",
"result": {
"burn_amount": 0,
"emission_amount": -10173000,
"fee_amount": 10173000,
"status": "OK"
}
}
2022-07-05 15:15:55 +02:00
burnt_oxen + = static_cast < int64_t > ( get_burned_amount_from_tx_extra ( tx . extra ) ) ;
2020-03-12 10:53:24 +01:00
}
2016-10-11 01:55:18 +02:00
}
2020-08-13 21:40:54 +02:00
2016-10-11 01:55:18 +02:00
emission_amount + = coinbase_amount - tx_fee_amount ;
total_fee_amount + = tx_fee_amount ;
2020-08-13 21:40:54 +02:00
if ( cache_to & & cache_to = = height )
{
std : : unique_lock lock { m_coinbase_cache . mutex } ;
2020-08-17 07:40:51 +02:00
if ( m_coinbase_cache . height < height )
{
m_coinbase_cache . height = height ;
m_coinbase_cache . emissions = emission_amount ;
m_coinbase_cache . fees = total_fee_amount ;
2021-01-04 01:09:45 +01:00
m_coinbase_cache . burnt = burnt_oxen ;
2020-08-17 07:40:51 +02:00
}
2020-08-13 21:40:54 +02:00
if ( m_coinbase_cache . building )
{
m_coinbase_cache . building = false ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Finishing cache build for get_coinbase_tx_sum in {} s " ,
std : : chrono : : duration < double > { std : : chrono : : steady_clock : : now ( ) - cache_build_started } . count ( ) ) ;
2020-08-13 21:40:54 +02:00
}
2020-08-17 07:40:51 +02:00
cache_to = 0 ;
2020-08-13 21:40:54 +02:00
}
2017-06-01 14:29:51 +02:00
return true ;
2020-08-13 21:40:54 +02:00
} ) ;
2016-10-10 21:45:51 +02:00
2020-08-13 21:40:54 +02:00
return result ;
2016-10-10 21:45:51 +02:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : check_tx_inputs_keyimages_diff ( const transaction & tx ) const
2014-03-03 23:07:58 +01:00
{
std : : unordered_set < crypto : : key_image > ki ;
2017-01-22 21:38:10 +01:00
for ( const auto & in : tx . vin )
2014-03-03 23:07:58 +01:00
{
2020-06-02 05:47:20 +02:00
CHECKED_GET_SPECIFIC_VARIANT ( in , txin_to_key , tokey_in , false ) ;
2014-03-03 23:07:58 +01:00
if ( ! ki . insert ( tokey_in . k_image ) . second )
return false ;
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2017-07-18 23:38:26 +02:00
bool core : : check_tx_inputs_ring_members_diff ( const transaction & tx ) const
{
2022-05-16 22:55:05 +02:00
const auto version = m_blockchain_storage . get_network_version ( ) ;
for ( const auto & in : tx . vin )
2017-07-18 23:38:26 +02:00
{
2022-05-16 22:55:05 +02:00
CHECKED_GET_SPECIFIC_VARIANT ( in , txin_to_key , tokey_in , false ) ;
for ( size_t n = 1 ; n < tokey_in . key_offsets . size ( ) ; + + n )
if ( tokey_in . key_offsets [ n ] = = 0 )
return false ;
2017-07-18 23:38:26 +02:00
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2017-02-20 18:41:39 +01:00
bool core : : check_tx_inputs_keyimages_domain ( const transaction & tx ) const
{
std : : unordered_set < crypto : : key_image > ki ;
for ( const auto & in : tx . vin )
{
2020-06-02 05:47:20 +02:00
CHECKED_GET_SPECIFIC_VARIANT ( in , txin_to_key , tokey_in , false ) ;
2017-02-20 18:41:39 +01:00
if ( ! ( rct : : scalarmultKey ( rct : : ki2rct ( tokey_in . k_image ) , rct : : curveOrder ( ) ) = = rct : : identity ( ) ) )
return false ;
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
size_t core : : get_blockchain_total_transactions ( ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_total_transactions ( ) ;
}
//-----------------------------------------------------------------------------------------------
2015-11-21 01:26:48 +01:00
bool core : : relay_txpool_transactions ( )
{
// we attempt to relay txes that should be relayed, but were not
2022-05-20 23:29:48 +02:00
std : : vector < std : : pair < crypto : : hash , std : : string > > txs ;
2017-03-12 15:51:32 +01:00
if ( m_mempool . get_relayable_transactions ( txs ) & & ! txs . empty ( ) )
2015-11-21 01:26:48 +01:00
{
2019-10-31 23:26:58 +01:00
cryptonote_connection_context fake_context { } ;
tx_verification_context tvc { } ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
NOTIFY_NEW_TRANSACTIONS : : request r { } ;
2015-11-21 01:26:48 +01:00
for ( auto it = txs . begin ( ) ; it ! = txs . end ( ) ; + + it )
{
2017-05-14 15:06:55 +02:00
r . txs . push_back ( it - > second ) ;
2015-11-21 01:26:48 +01:00
}
get_protocol ( ) - > relay_transactions ( r , fake_context ) ;
m_mempool . set_relayed ( txs ) ;
}
return true ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2018-07-12 03:58:17 +02:00
bool core : : submit_uptime_proof ( )
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( ! m_service_node )
2019-06-27 09:05:44 +02:00
return true ;
2018-09-27 10:31:08 +02:00
2019-10-31 23:26:58 +01:00
cryptonote_connection_context fake_context { } ;
2020-12-17 06:52:16 +01:00
bool relayed ;
auto height = get_current_blockchain_height ( ) ;
2021-04-30 02:23:16 +02:00
auto proof = m_service_node_list . generate_uptime_proof ( m_sn_public_ip , storage_https_port ( ) , storage_omq_port ( ) , ss_version , m_quorumnet_port , lokinet_version ) ;
NOTIFY_BTENCODED_UPTIME_PROOF : : request req = proof . generate_request ( ) ;
relayed = get_protocol ( ) - > relay_btencoded_uptime_proof ( req , fake_context ) ;
// TODO: remove after HF19
if ( relayed & & tools : : view_guts ( m_service_keys . pub ) ! = tools : : view_guts ( m_service_keys . pub_ed25519 ) ) {
// Temp workaround: nodes with both pub and ed25519 are failing bt-encoded proofs, so send
// an old-style proof out as well as a workaround.
2021-04-08 18:42:33 +02:00
NOTIFY_UPTIME_PROOF : : request req = m_service_node_list . generate_uptime_proof ( m_sn_public_ip , storage_https_port ( ) , storage_omq_port ( ) , m_quorumnet_port ) ;
2021-04-30 02:23:16 +02:00
get_protocol ( ) - > relay_uptime_proof ( req , fake_context ) ;
2020-12-17 06:52:16 +01:00
}
2021-04-30 02:23:16 +02:00
2019-06-27 09:05:44 +02:00
if ( relayed )
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Submitted uptime-proof for Service Node (yours): {} " , m_service_keys . pub ) ;
2019-06-07 09:03:11 +02:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2019-09-18 07:36:00 +02:00
bool core : : handle_uptime_proof ( const NOTIFY_UPTIME_PROOF : : request & proof , bool & my_uptime_proof_confirmation )
2019-06-07 09:03:11 +02:00
{
2020-08-06 09:21:00 +02:00
crypto : : x25519_public_key pkey = { } ;
bool result = m_service_node_list . handle_uptime_proof ( proof , my_uptime_proof_confirmation , pkey ) ;
if ( result & & m_service_node_list . is_service_node ( proof . pubkey , true /*require_active*/ ) & & pkey )
2020-08-02 08:20:11 +02:00
{
2021-01-14 21:28:50 +01:00
oxenmq : : pubkey_set added ;
2020-08-02 08:20:11 +02:00
added . insert ( tools : : copy_guts ( pkey ) ) ;
2021-04-08 18:38:51 +02:00
m_omq - > update_active_sns ( added , { } /*removed*/ ) ;
2020-08-02 08:20:11 +02:00
}
return result ;
2019-06-07 09:03:11 +02:00
}
//-----------------------------------------------------------------------------------------------
2020-12-17 06:52:16 +01:00
bool core : : handle_btencoded_uptime_proof ( const NOTIFY_BTENCODED_UPTIME_PROOF : : request & req , bool & my_uptime_proof_confirmation )
{
crypto : : x25519_public_key pkey = { } ;
2021-01-28 01:07:57 +01:00
auto proof = std : : make_unique < uptime_proof : : Proof > ( req . proof ) ;
proof - > sig = tools : : make_from_guts < crypto : : signature > ( req . sig ) ;
proof - > sig_ed25519 = tools : : make_from_guts < crypto : : ed25519_signature > ( req . ed_sig ) ;
auto pubkey = proof - > pubkey ;
bool result = m_service_node_list . handle_btencoded_uptime_proof ( std : : move ( proof ) , my_uptime_proof_confirmation , pkey ) ;
if ( result & & m_service_node_list . is_service_node ( pubkey , true /*require_active*/ ) & & pkey )
2020-12-17 06:52:16 +01:00
{
oxenmq : : pubkey_set added ;
added . insert ( tools : : copy_guts ( pkey ) ) ;
2021-04-08 18:38:51 +02:00
m_omq - > update_active_sns ( added , { } /*removed*/ ) ;
2020-12-17 06:52:16 +01:00
}
return result ;
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
crypto : : hash core : : on_transaction_relayed ( const std : : string & tx_blob )
2016-10-22 21:46:51 +02:00
{
2022-05-20 23:29:48 +02:00
std : : vector < std : : pair < crypto : : hash , std : : string > > txs ;
2016-10-22 21:46:51 +02:00
cryptonote : : transaction tx ;
2018-11-13 00:22:20 +01:00
crypto : : hash tx_hash ;
if ( ! parse_and_validate_tx_from_blob ( tx_blob , tx , tx_hash ) )
2016-10-22 21:46:51 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Failed to parse relayed transaction " ) ;
2019-11-06 07:28:33 +01:00
return crypto : : null_hash ;
2016-10-22 21:46:51 +02:00
}
2017-05-14 15:06:55 +02:00
txs . push_back ( std : : make_pair ( tx_hash , std : : move ( tx_blob ) ) ) ;
2016-10-22 21:46:51 +02:00
m_mempool . set_relayed ( txs ) ;
2019-11-06 07:28:33 +01:00
return tx_hash ;
2016-10-22 21:46:51 +02:00
}
//-----------------------------------------------------------------------------------------------
2019-05-31 03:06:42 +02:00
bool core : : relay_service_node_votes ( )
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
{
2019-10-15 19:45:58 +02:00
auto height = get_current_blockchain_height ( ) ;
2021-06-18 01:06:51 +02:00
auto hf_version = get_network_version ( m_nettype , height ) ;
2019-10-15 19:45:58 +02:00
2019-12-10 19:57:36 +01:00
auto quorum_votes = m_quorum_cop . get_relayable_votes ( height , hf_version , true ) ;
auto p2p_votes = m_quorum_cop . get_relayable_votes ( height , hf_version , false ) ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
if ( ! quorum_votes . empty ( ) & & m_quorumnet_state & & m_service_node )
quorumnet_relay_obligation_votes ( m_quorumnet_state , quorum_votes ) ;
2019-10-15 19:45:58 +02:00
2019-12-10 19:57:36 +01:00
if ( ! p2p_votes . empty ( ) )
{
2020-01-18 17:12:29 +01:00
NOTIFY_NEW_SERVICE_NODE_VOTE : : request req { } ;
req . votes = std : : move ( p2p_votes ) ;
cryptonote_connection_context fake_context { } ;
get_protocol ( ) - > relay_service_node_votes ( req , fake_context ) ;
2019-05-01 08:01:17 +02:00
}
return true ;
}
2019-12-10 19:57:36 +01:00
void core : : set_service_node_votes_relayed ( const std : : vector < service_nodes : : quorum_vote_t > & votes )
{
m_quorum_cop . set_votes_relayed ( votes ) ;
}
2019-05-01 08:01:17 +02:00
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : create_next_miner_block_template ( block & b , const account_public_address & adr , difficulty_type & diffic , uint64_t & height , uint64_t & expected_reward , const std : : string & ex_nonce )
2014-03-03 23:07:58 +01:00
{
2020-07-16 06:08:29 +02:00
return m_blockchain_storage . create_next_miner_block_template ( b , adr , diffic , height , expected_reward , ex_nonce ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : create_miner_block_template ( block & b , const crypto : : hash * prev_block , const account_public_address & adr , difficulty_type & diffic , uint64_t & height , uint64_t & expected_reward , const std : : string & ex_nonce )
2019-03-23 17:20:08 +01:00
{
2020-07-16 06:08:29 +02:00
return m_blockchain_storage . create_miner_block_template ( b , prev_block , adr , diffic , height , expected_reward , ex_nonce ) ;
2019-03-23 17:20:08 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : find_blockchain_supplement ( const std : : list < crypto : : hash > & qblock_ids , NOTIFY_RESPONSE_CHAIN_ENTRY : : request & resp ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . find_blockchain_supplement ( qblock_ids , resp ) ;
}
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : find_blockchain_supplement ( const uint64_t req_start_block , const std : : list < crypto : : hash > & qblock_ids , std : : vector < std : : pair < std : : pair < std : : string , crypto : : hash > , std : : vector < std : : pair < crypto : : hash , std : : string > > > > & blocks , uint64_t & total_height , uint64_t & start_height , bool pruned , bool get_miner_tx_hash , size_t max_count ) const
2014-03-03 23:07:58 +01:00
{
2018-04-16 21:16:00 +02:00
return m_blockchain_storage . find_blockchain_supplement ( req_start_block , qblock_ids , blocks , total_height , start_height , pruned , get_miner_tx_hash , max_count ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
bool core : : get_outs ( const rpc : : GET_OUTPUTS_BIN : : request & req , rpc : : GET_OUTPUTS_BIN : : response & res ) const
2016-08-02 22:48:09 +02:00
{
return m_blockchain_storage . get_outs ( req , res ) ;
}
//-----------------------------------------------------------------------------------------------
2018-05-16 23:41:50 +02:00
bool core : : get_output_distribution ( uint64_t amount , uint64_t from_height , uint64_t to_height , uint64_t & start_height , std : : vector < uint64_t > & distribution , uint64_t & base ) const
2018-02-19 12:15:15 +01:00
{
2018-05-16 23:41:50 +02:00
return m_blockchain_storage . get_output_distribution ( amount , from_height , to_height , start_height , distribution , base ) ;
2018-02-19 12:15:15 +01:00
}
//-----------------------------------------------------------------------------------------------
2020-06-02 07:50:27 +02:00
void core : : get_output_blacklist ( std : : vector < uint64_t > & blacklist ) const
Infinite Staking Part 2 (#406)
* Cleanup and undoing some protocol breakages
* Simplify expiration of nodes
* Request unlock schedules entire node for expiration
* Fix off by one in expiring nodes
* Undo expiring code for pre v10 nodes
* Fix RPC returning register as unlock height and not checking 0
* Rename key image unlock height const
* Undo testnet hardfork debug changes
* Remove is_type for get_type, fix missing var rename
* Move serialisable data into public namespace
* Serialise tx types properly
* Fix typo in no service node known msg
* Code review
* Fix == to >= on serialising tx type
* Code review 2
* Fix tests and key image unlock
* Add command to print locked key images
* Update ui to display lock stakes, query in print cmd blacklist
* Modify print stakes to be less slow
* Remove autostaking code
* Refactor staking into sweep functions
It appears staking was derived off stake_main written separately at
implementation at the beginning. This merges them back into a common
code path, after removing autostake there's only some minor differences.
It also makes sure that any changes to sweeping upstream are going to be
considered in the staking process which we want.
* Display unlock height for stakes
* Begin creating output blacklist
* Make blacklist output a migration step
* Implement get_output_blacklist for lmdb
* In wallet output selection ignore blacklisted outputs
* Apply blacklisted outputs to output selection
* Fix broken tests, switch key image unlock
* Fix broken unit_tests
* Begin change to limit locked key images to 4 globally
* Revamp prepare registration for new min contribution rules
* Fix up old back case in prepare registration
* Remove debug code
* Cleanup debug code and some unecessary changes
* Fix migration step on mainnet db
* Fix blacklist outputs for pre-existing DB's
* Remove irrelevant note
* Tweak scanning addresses for locked stakes
Since we only now allow contributions from the primary address we can
skip checking all subaddress + lookahead to speed up wallet scanning
* Define macro for SCNu64 for Mingw
* Fix failure on empty DB
* Add missing error msg, remove contributor from stake
* Improve staking messages
* Flush prompt to always display
* Return the msg from stake failure and fix stake parsing error
* Tweak fork rules for smaller bulletproofs
* Tweak pooled nodes minimum amounts
* Fix crash on exit, there's no need to store on destructor
Since all information about service nodes is derived from the blockchain
and we store state every time we receive a block, storing in the
destructor is redundant as there is no new information to store.
* Make prompt be consistent with CLI
* Check max number of key images from per user to node
* Implement error message on get_output_blacklist failure
* Remove resolved TODO's/comments
* Handle infinite staking in print_sn
* Atoi->strtol, fix prepare_registration, virtual override, stale msgs
2019-02-14 02:12:57 +01:00
{
2020-06-02 07:50:27 +02:00
m_blockchain_storage . get_output_blacklist ( blacklist ) ;
Infinite Staking Part 2 (#406)
* Cleanup and undoing some protocol breakages
* Simplify expiration of nodes
* Request unlock schedules entire node for expiration
* Fix off by one in expiring nodes
* Undo expiring code for pre v10 nodes
* Fix RPC returning register as unlock height and not checking 0
* Rename key image unlock height const
* Undo testnet hardfork debug changes
* Remove is_type for get_type, fix missing var rename
* Move serialisable data into public namespace
* Serialise tx types properly
* Fix typo in no service node known msg
* Code review
* Fix == to >= on serialising tx type
* Code review 2
* Fix tests and key image unlock
* Add command to print locked key images
* Update ui to display lock stakes, query in print cmd blacklist
* Modify print stakes to be less slow
* Remove autostaking code
* Refactor staking into sweep functions
It appears staking was derived off stake_main written separately at
implementation at the beginning. This merges them back into a common
code path, after removing autostake there's only some minor differences.
It also makes sure that any changes to sweeping upstream are going to be
considered in the staking process which we want.
* Display unlock height for stakes
* Begin creating output blacklist
* Make blacklist output a migration step
* Implement get_output_blacklist for lmdb
* In wallet output selection ignore blacklisted outputs
* Apply blacklisted outputs to output selection
* Fix broken tests, switch key image unlock
* Fix broken unit_tests
* Begin change to limit locked key images to 4 globally
* Revamp prepare registration for new min contribution rules
* Fix up old back case in prepare registration
* Remove debug code
* Cleanup debug code and some unecessary changes
* Fix migration step on mainnet db
* Fix blacklist outputs for pre-existing DB's
* Remove irrelevant note
* Tweak scanning addresses for locked stakes
Since we only now allow contributions from the primary address we can
skip checking all subaddress + lookahead to speed up wallet scanning
* Define macro for SCNu64 for Mingw
* Fix failure on empty DB
* Add missing error msg, remove contributor from stake
* Improve staking messages
* Flush prompt to always display
* Return the msg from stake failure and fix stake parsing error
* Tweak fork rules for smaller bulletproofs
* Tweak pooled nodes minimum amounts
* Fix crash on exit, there's no need to store on destructor
Since all information about service nodes is derived from the blockchain
and we store state every time we receive a block, storing in the
destructor is redundant as there is no new information to store.
* Make prompt be consistent with CLI
* Check max number of key images from per user to node
* Implement error message on get_output_blacklist failure
* Remove resolved TODO's/comments
* Handle infinite staking in print_sn
* Atoi->strtol, fix prepare_registration, virtual override, stale msgs
2019-02-14 02:12:57 +01:00
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : get_tx_outputs_gindexs ( const crypto : : hash & tx_id , std : : vector < uint64_t > & indexs ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_tx_outputs_gindexs ( tx_id , indexs ) ;
}
//-----------------------------------------------------------------------------------------------
2018-12-16 14:28:49 +01:00
bool core : : get_tx_outputs_gindexs ( const crypto : : hash & tx_id , size_t n_txes , std : : vector < std : : vector < uint64_t > > & indexs ) const
{
return m_blockchain_storage . get_tx_outputs_gindexs ( tx_id , n_txes , indexs ) ;
}
//-----------------------------------------------------------------------------------------------
2014-03-03 23:07:58 +01:00
void core : : pause_mine ( )
{
m_miner . pause ( ) ;
}
//-----------------------------------------------------------------------------------------------
void core : : resume_mine ( )
{
m_miner . resume ( ) ;
}
//-----------------------------------------------------------------------------------------------
2017-03-19 11:25:12 +01:00
block_complete_entry get_block_complete_entry ( block & b , tx_memory_pool & pool )
{
2020-05-25 03:48:32 +02:00
block_complete_entry bce = { } ;
bce . block = cryptonote : : block_to_blob ( b ) ;
2017-03-19 11:25:12 +01:00
for ( const auto & tx_hash : b . tx_hashes )
{
2022-05-20 23:29:48 +02:00
std : : string txblob ;
2017-05-14 15:06:55 +02:00
CHECK_AND_ASSERT_THROW_MES ( pool . get_transaction ( tx_hash , txblob ) , " Transaction not found in pool " ) ;
bce . txs . push_back ( txblob ) ;
2017-03-19 11:25:12 +01:00
}
return bce ;
}
//-----------------------------------------------------------------------------------------------
2019-03-23 17:20:08 +01:00
bool core : : handle_block_found ( block & b , block_verification_context & bvc )
2014-03-03 23:07:58 +01:00
{
2019-10-31 23:26:58 +01:00
bvc = { } ;
2018-04-16 01:16:02 +02:00
std : : vector < block_complete_entry > blocks ;
2019-08-12 01:48:47 +02:00
m_miner . pause ( ) ;
2018-11-23 14:05:48 +01:00
{
2021-01-04 04:19:42 +01:00
OXEN_DEFER { m_miner . resume ( ) ; } ;
2019-08-12 01:48:47 +02:00
try
{
blocks . push_back ( get_block_complete_entry ( b , m_mempool ) ) ;
}
catch ( const std : : exception & e )
{
return false ;
}
std : : vector < block > pblocks ;
if ( ! prepare_handle_incoming_blocks ( blocks , pblocks ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( logcat , " Block found, but failed to prepare to add " ) ;
2019-08-12 01:48:47 +02:00
return false ;
}
2022-04-29 01:51:14 +02:00
// add_new_block will verify block and set bvc.m_verification_failed accordingly
2019-08-12 01:48:47 +02:00
add_new_block ( b , bvc , nullptr /*checkpoint*/ ) ;
cleanup_handle_incoming_blocks ( true ) ;
m_miner . on_block_chain_update ( ) ;
2018-11-23 14:05:48 +01:00
}
2014-03-03 23:07:58 +01:00
2020-09-24 07:16:21 +02:00
if ( bvc . m_verifivation_failed )
{
bool pulse = cryptonote : : block_has_pulse_components ( b ) ;
2022-09-13 08:57:56 +02:00
oxen : : log : : error ( oxen : : log : : Cat ( " verify " ) , " {} block failed verification \n {} " , ( pulse ? " Pulse " : " Mined " ) , cryptonote : : obj_to_json_str ( b ) ) ;
2020-09-24 07:16:21 +02:00
return false ;
}
else if ( bvc . m_added_to_main_chain )
2014-03-03 23:07:58 +01:00
{
2021-08-19 05:43:26 +02:00
std : : unordered_set < crypto : : hash > missed_txs ;
2022-05-20 23:29:48 +02:00
std : : vector < std : : string > txs ;
2021-08-19 05:43:26 +02:00
m_blockchain_storage . get_transactions_blobs ( b . tx_hashes , txs , & missed_txs ) ;
2014-03-03 23:07:58 +01:00
if ( missed_txs . size ( ) & & m_blockchain_storage . get_block_id_by_height ( get_block_height ( b ) ) ! = get_block_hash ( b ) )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Block found but, seems that reorganize just happened after that, do not relay this block " ) ;
2014-03-03 23:07:58 +01:00
return true ;
}
2018-03-01 12:36:19 +01:00
CHECK_AND_ASSERT_MES ( txs . size ( ) = = b . tx_hashes . size ( ) & & ! missed_txs . size ( ) , false , " can't find some transactions in found block: " < < get_block_hash ( b ) < < " txs.size()= " < < txs . size ( )
2014-03-03 23:07:58 +01:00
< < " , b.tx_hashes.size()= " < < b . tx_hashes . size ( ) < < " , missed_txs.size() " < < missed_txs . size ( ) ) ;
2019-10-31 23:26:58 +01:00
cryptonote_connection_context exclude_context { } ;
NOTIFY_NEW_FLUFFY_BLOCK : : request arg { } ;
2019-06-04 04:28:56 +02:00
arg . current_blockchain_height = m_blockchain_storage . get_current_blockchain_height ( ) ;
arg . b = blocks [ 0 ] ;
2014-03-03 23:07:58 +01:00
m_pprotocol - > relay_block ( arg , exclude_context ) ;
}
2019-03-23 17:20:08 +01:00
return true ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
void core : : on_synchronized ( )
{
m_miner . on_synchronized ( ) ;
}
//-----------------------------------------------------------------------------------------------
2017-08-22 15:54:18 +02:00
void core : : safesyncmode ( const bool onoff )
{
m_blockchain_storage . safesyncmode ( onoff ) ;
}
//-----------------------------------------------------------------------------------------------
2019-06-26 06:00:05 +02:00
bool core : : add_new_block ( const block & b , block_verification_context & bvc , checkpoint_t const * checkpoint )
2014-03-03 23:07:58 +01:00
{
2019-06-26 06:00:05 +02:00
bool result = m_blockchain_storage . add_new_block ( b , bvc , checkpoint ) ;
if ( result )
relay_service_node_votes ( ) ; // NOTE: nop if synchronising due to not accepting votes whilst syncing
return result ;
2014-03-03 23:07:58 +01:00
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
//-----------------------------------------------------------------------------------------------
2019-06-26 06:00:05 +02:00
bool core : : prepare_handle_incoming_blocks ( const std : : vector < block_complete_entry > & blocks_entry , std : : vector < block > & blocks )
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
{
2017-05-14 15:06:55 +02:00
m_incoming_tx_lock . lock ( ) ;
2019-06-26 06:00:05 +02:00
if ( ! m_blockchain_storage . prepare_handle_incoming_blocks ( blocks_entry , blocks ) )
2018-11-23 14:05:48 +01:00
{
cleanup_handle_incoming_blocks ( false ) ;
return false ;
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
return true ;
}
//-----------------------------------------------------------------------------------------------
bool core : : cleanup_handle_incoming_blocks ( bool force_sync )
{
2017-08-29 16:43:32 +02:00
bool success = false ;
2017-05-14 15:06:55 +02:00
try {
2017-08-29 16:43:32 +02:00
success = m_blockchain_storage . cleanup_handle_incoming_blocks ( force_sync ) ;
2017-05-14 15:06:55 +02:00
}
catch ( . . . ) { }
m_incoming_tx_lock . unlock ( ) ;
2017-08-29 16:43:32 +02:00
return success ;
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
}
2014-03-03 23:07:58 +01:00
//-----------------------------------------------------------------------------------------------
2022-05-20 23:29:48 +02:00
bool core : : handle_incoming_block ( const std : : string & block_blob , const block * b , block_verification_context & bvc , checkpoint_t * checkpoint , bool update_miner_blocktemplate )
2014-03-03 23:07:58 +01:00
{
2017-08-29 14:01:45 +02:00
TRY_ENTRY ( ) ;
2019-10-31 23:26:58 +01:00
bvc = { } ;
2019-04-04 02:15:57 +02:00
if ( ! check_incoming_block_size ( block_blob ) )
2014-03-03 23:07:58 +01:00
{
bvc . m_verifivation_failed = true ;
return false ;
}
2019-04-04 02:15:57 +02:00
if ( ( ( size_t ) - 1 ) < = 0xffffffff & & block_blob . size ( ) > = 0x3fffffff )
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , " This block's size is {}, closing on the 32 bit limit " , block_blob . size ( ) ) ;
2019-04-04 02:15:57 +02:00
2019-07-11 06:16:55 +02:00
CHECK_AND_ASSERT_MES ( update_checkpoints_from_json_file ( ) , false , " One or more checkpoints loaded from json conflicted with existing checkpoints. " ) ;
2019-04-04 02:15:57 +02:00
2018-11-23 12:15:04 +01:00
block lb ;
if ( ! b )
2014-03-03 23:07:58 +01:00
{
2018-12-13 20:47:47 +01:00
crypto : : hash block_hash ;
if ( ! parse_and_validate_block_from_blob ( block_blob , lb , block_hash ) )
2018-11-23 12:15:04 +01:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " Failed to parse and validate new block " ) ;
2018-11-23 12:15:04 +01:00
bvc . m_verifivation_failed = true ;
return false ;
}
b = & lb ;
2014-03-03 23:07:58 +01:00
}
2019-08-14 06:52:51 +02:00
2019-06-26 06:00:05 +02:00
add_new_block ( * b , bvc , checkpoint ) ;
2014-03-03 23:07:58 +01:00
if ( update_miner_blocktemplate & & bvc . m_added_to_main_chain )
2019-08-07 02:30:46 +02:00
m_miner . on_block_chain_update ( ) ;
2014-03-03 23:07:58 +01:00
return true ;
2017-08-29 14:01:45 +02:00
CATCH_ENTRY_L0 ( " core::handle_incoming_block() " , false ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2014-06-11 23:32:53 +02:00
// Used by the RPC server to check the size of an incoming
// block_blob
2022-05-20 23:29:48 +02:00
bool core : : check_incoming_block_size ( const std : : string & block_blob ) const
2014-06-11 23:32:53 +02:00
{
2019-04-04 02:15:57 +02:00
// note: we assume block weight is always >= block blob size, so we check incoming
// blob size against the block weight limit, which acts as a sanity check without
// having to parse/weigh first; in fact, since the block blob is the block header
// plus the tx hashes, the weight will typically be much larger than the blob size
if ( block_blob . size ( ) > m_blockchain_storage . get_current_cumulative_block_weight_limit ( ) + BLOCK_SIZE_SANITY_LEEWAY )
2014-06-11 23:32:53 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , " WRONG BLOCK BLOB, sanity check failed on size {}, rejected " , block_blob . size ( ) ) ;
2014-06-11 23:32:53 +02:00
return false ;
}
return true ;
}
2020-04-15 01:22:18 +02:00
2021-04-08 18:38:51 +02:00
void core : : update_omq_sns ( )
2020-04-15 01:22:18 +02:00
{
2021-01-04 05:21:21 +01:00
// TODO: let callers (e.g. lokinet, ss) subscribe to callbacks when this fires
2021-01-14 21:28:50 +01:00
oxenmq : : pubkey_set active_sns ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
m_service_node_list . copy_active_x25519_pubkeys ( std : : inserter ( active_sns , active_sns . end ( ) ) ) ;
2021-04-08 18:38:51 +02:00
m_omq - > set_active_sns ( std : : move ( active_sns ) ) ;
2020-04-15 01:22:18 +02:00
}
2014-06-11 23:32:53 +02:00
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
crypto : : hash core : : get_tail_id ( ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_tail_id ( ) ;
}
//-----------------------------------------------------------------------------------------------
2017-08-12 16:24:39 +02:00
difficulty_type core : : get_block_cumulative_difficulty ( uint64_t height ) const
{
return m_blockchain_storage . get_db ( ) . get_block_cumulative_difficulty ( height ) ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
bool core : : have_block ( const crypto : : hash & id ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . have_block ( id ) ;
}
//-----------------------------------------------------------------------------------------------
2015-09-19 12:25:57 +02:00
crypto : : hash core : : get_block_id_by_height ( uint64_t height ) const
2014-03-03 23:07:58 +01:00
{
return m_blockchain_storage . get_block_id_by_height ( height ) ;
}
//-----------------------------------------------------------------------------------------------
2017-01-22 13:20:55 +01:00
bool core : : get_block_by_hash ( const crypto : : hash & h , block & blk , bool * orphan ) const
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
{
2017-01-22 13:20:55 +01:00
return m_blockchain_storage . get_block_by_hash ( h , blk , orphan ) ;
2014-03-03 23:07:58 +01:00
}
//-----------------------------------------------------------------------------------------------
2020-08-28 21:11:25 +02:00
bool core : : get_block_by_height ( uint64_t height , block & blk ) const
{
return m_blockchain_storage . get_block_by_height ( height , blk ) ;
}
//-----------------------------------------------------------------------------------------------
2021-03-24 23:20:16 +01:00
static bool check_external_ping ( time_t last_ping , std : : chrono : : seconds lifetime , std : : string_view what )
2019-06-27 09:05:44 +02:00
{
2021-03-24 23:20:16 +01:00
const std : : chrono : : seconds elapsed { std : : time ( nullptr ) - last_ping } ;
2019-11-26 23:43:15 +01:00
if ( elapsed > lifetime )
2019-06-27 09:05:44 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , " Have not heard from {} {} " , what ,
2019-11-26 23:43:15 +01:00
( ! last_ping ? " since starting " :
2021-03-24 23:20:16 +01:00
" since more than " + tools : : get_human_readable_timespan ( elapsed ) + " ago " ) ) ;
2019-06-27 09:05:44 +02:00
return false ;
}
return true ;
}
2019-12-20 19:21:24 +01:00
void core : : reset_proof_interval ( )
{
m_check_uptime_proof_interval . reset ( ) ;
}
2019-06-27 09:05:44 +02:00
//-----------------------------------------------------------------------------------------------
2018-08-22 12:47:32 +02:00
void core : : do_uptime_proof_call ( )
{
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
std : : vector < service_nodes : : service_node_pubkey_info > const states = get_service_node_list_state ( { m_service_keys . pub } ) ;
2019-06-27 09:05:44 +02:00
2022-04-29 01:51:14 +02:00
// wait one block before starting uptime proofs (but not on testnet/devnet, where we sometimes
// have mass registrations/deregistrations where the waiting causes problems).
2022-05-16 22:55:05 +02:00
uint64_t delay_blocks = m_nettype = = network_type : : MAINNET ? 1 : 0 ;
2022-04-29 01:51:14 +02:00
if ( ! states . empty ( ) & & ( states [ 0 ] . info - > registration_height + delay_blocks ) < get_current_blockchain_height ( ) )
2018-08-22 12:47:32 +02:00
{
2019-12-04 21:50:28 +01:00
m_check_uptime_proof_interval . do_call ( [ this ] ( ) {
2019-10-27 23:10:22 +01:00
// This timer is not perfectly precise and can leak seconds slightly, so send the uptime
// proof if we are within half a tick of the target time. (Essentially our target proof
2021-03-24 23:20:16 +01:00
// window becomes the first time this triggers in the 59.75-60.25 minute window).
2019-12-04 21:50:28 +01:00
uint64_t next_proof_time = 0 ;
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
m_service_node_list . access_proof ( m_service_keys . pub , [ & ] ( auto & proof ) { next_proof_time = proof . timestamp ; } ) ;
2021-03-24 23:20:16 +01:00
auto & netconf = get_net_config ( ) ;
next_proof_time + = std : : chrono : : seconds {
netconf . UPTIME_PROOF_FREQUENCY - netconf . UPTIME_PROOF_CHECK_INTERVAL / 2 } . count ( ) ;
2019-12-04 21:50:28 +01:00
if ( ( uint64_t ) std : : time ( nullptr ) < next_proof_time )
2019-12-20 19:21:24 +01:00
return ;
2019-12-04 21:50:28 +01:00
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
auto pubkey = m_service_node_list . get_pubkey_from_x25519 ( m_service_keys . pub_x25519 ) ;
2020-10-19 00:38:21 +02:00
if ( pubkey ! = crypto : : null_pkey & & pubkey ! = m_service_keys . pub & & m_service_node_list . is_service_node ( pubkey , false /*don't require active*/ ) )
2020-04-14 02:53:40 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : red ) ,
" Failed to submit uptime proof: another service node on the network is using the same ed/x25519 keys as this service node. This typically means both have the same 'key_ed25519' private key file. " ) ) ;
2020-04-14 02:53:40 +02:00
return ;
}
2020-10-19 16:33:13 +02:00
{
std : : vector < crypto : : public_key > sn_pks ;
auto sns = m_service_node_list . get_service_node_list_state ( ) ;
sn_pks . reserve ( sns . size ( ) ) ;
for ( const auto & sni : sns )
sn_pks . push_back ( sni . pubkey ) ;
m_service_node_list . for_each_service_node_info_and_proof ( sn_pks . begin ( ) , sn_pks . end ( ) , [ & ] ( auto & pk , auto & sni , auto & proof ) {
2021-01-28 01:07:57 +01:00
if ( pk ! = m_service_keys . pub & & proof . proof - > public_ip = = m_sn_public_ip & &
2022-05-25 19:40:38 +02:00
( proof . proof - > qnet_port = = m_quorumnet_port | | (
m_nettype ! = network_type : : DEVNET & & ( proof . proof - > storage_https_port = = storage_https_port ( ) | | proof . proof - > storage_omq_port = = storage_omq_port ( ) ) ) ) )
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : red ) , " Another service node ({}) is broadcasting the same public IP and ports as this service node ({}:{}[qnet], :{}[SS-HTTP], :{}[SS-LMQ]). This will lead to deregistration of one or both service nodes if not corrected. (Do both service nodes have the correct IP for the service-node-public-ip setting?) " , pk , epee : : string_tools : : get_ip_string_from_int32 ( m_sn_public_ip ) , proof . proof - > qnet_port , proof . proof - > storage_https_port , proof . proof - > storage_omq_port ) ) ;
2020-10-19 16:33:13 +02:00
} ) ;
}
2022-05-16 22:55:05 +02:00
if ( m_nettype ! = network_type : : DEVNET )
2019-06-27 09:05:44 +02:00
{
2021-03-24 23:20:16 +01:00
if ( ! check_external_ping ( m_last_storage_server_ping , get_net_config ( ) . UPTIME_PROOF_FREQUENCY , " the storage server " ) )
2020-08-12 00:00:42 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : red ) ,
" Failed to submit uptime proof: have not heard from the storage server recently. Make sure that it is running! It is required to run alongside the Loki daemon " ) ) ;
2020-08-12 00:00:42 +02:00
return ;
}
2021-03-24 23:20:16 +01:00
if ( ! check_external_ping ( m_last_lokinet_ping , get_net_config ( ) . UPTIME_PROOF_FREQUENCY , " Lokinet " ) )
2019-06-27 09:05:44 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : red ) ,
" Failed to submit uptime proof: have not heard from lokinet recently. Make sure that it is running! It is required to run alongside the Loki daemon " ) ) ;
2019-12-20 19:21:24 +01:00
return ;
2019-06-07 09:03:11 +02:00
}
}
2018-09-27 10:31:08 +02:00
2019-12-04 21:50:28 +01:00
submit_uptime_proof ( ) ;
2018-09-27 10:31:08 +02:00
} ) ;
2018-08-22 12:47:32 +02:00
}
else
{
2018-09-27 10:31:08 +02:00
// reset the interval so that we're ready when we register, OR if we get deregistered this primes us up for re-registration in the same session
2019-12-20 19:21:24 +01:00
m_check_uptime_proof_interval . reset ( ) ;
2018-08-22 12:47:32 +02:00
}
}
//-----------------------------------------------------------------------------------------------
2014-03-03 23:07:58 +01:00
bool core : : on_idle ( )
{
if ( ! m_starter_message_showed )
{
2017-11-30 16:35:52 +01:00
std : : string main_message ;
if ( m_offline )
2018-04-10 06:49:20 +02:00
main_message = " The daemon is running offline and will not attempt to sync to the Loki network. " ;
2017-11-30 16:35:52 +01:00
else
main_message = " The daemon will start synchronizing with the network. This may take a long time to complete. " ;
2022-09-13 08:57:56 +02:00
oxen : : log : : info ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : yellow ) , " \n ********************************************************************** \n \
{ } \ n \ n \
You can set the level of process detailization through \ " set_log <level|categories> \" command, \n \
where < level > is between 0 ( no details ) and 4 ( very verbose ) , or custom category based levels ( eg , * : WARNING ) . \ n \
\ n \
Use the \ " help \" command to see the list of available commands. \n \
Use \ " help <command> \" to see a command's documentation. \n \
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * \ n " , main_message));
2014-03-03 23:07:58 +01:00
m_starter_message_showed = true ;
}
2020-06-03 05:10:08 +02:00
m_txpool_auto_relayer . do_call ( [ this ] { return relay_txpool_transactions ( ) ; } ) ;
m_service_node_vote_relayer . do_call ( [ this ] { return relay_service_node_votes ( ) ; } ) ;
m_check_disk_space_interval . do_call ( [ this ] { return check_disk_space ( ) ; } ) ;
m_block_rate_interval . do_call ( [ this ] { return check_block_rate ( ) ; } ) ;
2019-12-04 21:50:28 +01:00
m_sn_proof_cleanup_interval . do_call ( [ & snl = m_service_node_list ] { snl . cleanup_proofs ( ) ; return true ; } ) ;
2018-09-27 10:31:08 +02:00
2021-03-24 23:20:16 +01:00
std : : chrono : : seconds lifetime { time ( nullptr ) - get_start_time ( ) } ;
if ( m_service_node & & lifetime > get_net_config ( ) . UPTIME_PROOF_STARTUP_DELAY ) // Give us some time to connect to peers before sending uptimes
2018-09-27 10:31:08 +02:00
{
2018-08-22 12:47:32 +02:00
do_uptime_proof_call ( ) ;
2018-09-27 10:31:08 +02:00
}
2020-06-03 05:10:08 +02:00
m_blockchain_pruning_interval . do_call ( [ this ] { return update_blockchain_pruning ( ) ; } ) ;
2014-03-03 23:07:58 +01:00
m_miner . on_idle ( ) ;
2014-06-15 09:48:13 +02:00
m_mempool . on_idle ( ) ;
2019-01-15 10:09:17 +01:00
2020-01-19 23:44:18 +01:00
# ifdef ENABLE_SYSTEMD
RPC overhaul
High-level details:
This redesigns the RPC layer to make it much easier to work with,
decouples it from an embedded HTTP server, and gets the vast majority of
the RPC serialization and dispatch code out of a very commonly included
header.
There is unfortunately rather a lot of interconnected code here that
cannot be easily separated out into separate commits. The full details
of what happens here are as follows:
Major details:
- All of the RPC code is now in a `cryptonote::rpc` namespace; this
renames quite a bit to be less verbose: e.g. CORE_RPC_STATUS_OK
becomes `rpc::STATUS_OK`, and `cryptonote::COMMAND_RPC_SOME_LONG_NAME`
becomes `rpc::SOME_LONG_NAME` (or just SOME_LONG_NAME for code already
working in the `rpc` namespace).
- `core_rpc_server` is now completely decoupled from providing any
request protocol: it is now *just* the core RPC call handler.
- The HTTP RPC interface now lives in a new rpc/http_server.h; this code
handles listening for HTTP requests and dispatching them to
core_rpc_server, then sending the results back to the caller.
- There is similarly a rpc/lmq_server.h for LMQ RPC code; more details
on this (and other LMQ specifics) below.
- RPC implementing code now returns the response object and throws when
things go wrong which simplifies much of the rpc error handling. They
can throw anything; generic exceptions get logged and a generic
"internal error" message gets returned to the caller, but there is
also an `rpc_error` class to return an error code and message used by
some json-rpc commands.
- RPC implementing functions now overload `core_rpc_server::invoke`
following the pattern:
RPC_BLAH_BLAH::response core_rpc_server::invoke(RPC_BLAH_BLAH::request&& req, rpc_context context);
This overloading makes the code vastly simpler: all instantiations are
now done with a small amount of generic instantiation code in a single
.cpp rather than needing to go to hell and back with a nest of epee
macros in a core header.
- each RPC endpoint is now defined by the RPC types themselves,
including its accessible names and permissions, in
core_rpc_server_commands_defs.h:
- every RPC structure now has a static `names()` function that returns
the names by which the end point is accessible. (The first one is
the primary, the others are for deprecated aliases).
- RPC command wrappers define their permissions and type by inheriting
from special tag classes:
- rpc::RPC_COMMAND is a basic, admin-only, JSON command, available
via JSON RPC. *All* JSON commands are now available via JSON RPC,
instead of the previous mix of some being at /foo and others at
/json_rpc. (Ones that were previously at /foo are still there for
backwards compatibility; see `rpc::LEGACY` below).
- rpc::PUBLIC specifies that the command should be available via a
restricted RPC connection.
- rpc::BINARY specifies that the command is not JSON, but rather is
accessible as /name and takes and returns values in the magic epee
binary "portable storage" (lol) data format.
- rpc::LEGACY specifies that the command should be available via the
non-json-rpc interface at `/name` for backwards compatibility (in
addition to the JSON-RPC interface).
- some epee serialization got unwrapped and de-templatized so that it
can be moved into a .cpp file with just declarations in the .h. (This
makes a *huge* difference for core_rpc_server_commands_defs.h and for
every compilation unit that includes it which previously had to
compile all the serialization code and then throw all by one copy away
at link time). This required some new macros so as to not break a ton
of places that will use the old way putting everything in the headers;
The RPC code uses this as does a few other places; there are comments
in contrib/epee/include/serialization/keyvalue_serialization.h as to
how to use it.
- Detemplatized a bunch of epee/storages code. Most of it should have
have been using templates at all (because it can only ever be called
with one type!), and now it isn't. This broke some things that didn't
properly compile because of missing headers or (in one case) a messed
up circular dependency.
- Significantly simplified a bunch of over-templatized serialization
code.
- All RPC serialization definitions is now out of
core_rpc_server_commands_defs.h and into a single .cpp file
(core_rpc_server_commands_defs.cpp).
- core RPC no longer uses the disgusting
BEGIN_URI_MAP2/MAP_URI_BLAH_BLAH macros. This was a terrible design
that forced slamming tons of code into a common header that didn't
need to be there.
- epee::struct_init is gone. It was a horrible hack that instiated
multiple templates just so the coder could be so lazy and write
`some_type var;` instead of properly value initializing with
`some_type var{};`.
- Removed a bunch of useless crap from epee. In particular, forcing
extra template instantiations all over the place in order to nest
return objects inside JSON RPC values is no longer needed, as are a
bunch of stuff related to the above de-macroization of the code.
- get_all_service_nodes, get_service_nodes, and get_n_service_nodes are
now combined into a single `get_service_nodes` (with deprecated
aliases for the others), which eliminates a fair amount of
duplication. The biggest obstacle here was getting the requested
fields reference passed through: this is now done by a new ability to
stash a context in the serialization object that can be retrieved by a
sub-serialized type.
LMQ-specifics:
- The LokiMQ instance moves into `cryptonote::core` rather than being
inside cryptonote_protocol. Currently the instance is used both for
qnet and rpc calls (and so needs to be in a common place), but I also
intend future PRs to use the batching code for job processing
(replacing the current threaded job queue).
- rpc/lmq_server.h handles the actual LMQ-request-to-core-RPC glue.
Unlike http_server it isn't technically running the whole LMQ stack
from here, but the parallel name with http_server seemed appropriate.
- All RPC endpoints are supported by LMQ under the same names as defined
generically, but prefixed with `rpc.` for public commands and `admin.`
for restricted ones.
- service node keys are now always available, even when not running in
`--service-node` mode: this is because we want the x25519 key for
being able to offer CURVE encryption for lmq RPC end-points, and
because it doesn't hurt to have them available all the time. In the
RPC layer this is now called "get_service_keys" (with
"get_service_node_key" as an alias) since they aren't strictly only
for service nodes. This also means code needs to check
m_service_node, and not m_service_node_keys, to tell if it is running
as a service node. (This is also easier to notice because
m_service_node_keys got renamed to `m_service_keys`).
- Added block and mempool monitoring LMQ RPC endpoints: `sub.block` and
`sub.mempool` subscribes the connection for new block and new mempool
TX notifications. The latter can notify on just blink txes, or all
new mempool txes (but only new ones -- txes dumped from a block don't
trigger it). The client gets pushed a [`notify.block`, `height`,
`hash`] or [`notify.tx`, `txhash`, `blob`] message when something
arrives.
Minor details:
- rpc::version_t is now a {major,minor} pair. Forcing everyone to pack
and unpack a uint32_t was gross.
- Changed some macros to constexprs (e.g. CORE_RPC_ERROR_CODE_...).
(This immediately revealed a couple of bugs in the RPC code that was
assigning CORE_RPC_ERROR_CODE_... to a string, and it worked because
the macro allows implicit conversion to a char).
- De-templatizing useless templates in epee (i.e. a bunch of templated
types that were never invoked with different types) revealed a painful
circular dependency between epee and non-epee code for tor_address and
i2p_address. This crap is now handled in a suitably named
`net/epee_network_address_hack.cpp` hack because it really isn't
trivial to extricate this mess.
- Removed `epee/include/serialization/serialize_base.h`. Amazingly the
code somehow still all works perfectly with this previously vital
header removed.
- Removed bitrotted, unused epee "crypted_storage" and
"gzipped_inmemstorage" code.
- Replaced a bunch of epee::misc_utils::auto_scope_leave_caller with
LOKI_DEFERs. The epee version involves quite a bit more instantiation
and is ugly as sin. Also made the `loki::defer` class invokable for
some edge cases that need calling before destruction in particular
conditions.
- Moved the systemd code around; it makes much more sense to do the
systemd started notification as in daemon.cpp as late as possible
rather than in core (when we can still have startup failures, e.g. if
the RPC layer can't start).
- Made the systemd short status string available in the get_info RPC
(and no longer require building with systemd).
- during startup, print (only) the x25519 when not in SN mode, and
continue to print all three when in SN mode.
- DRYed out some RPC implementation code (such as set_limit)
- Made wallet_rpc stop using a raw m_wallet pointer
2020-04-28 01:25:43 +02:00
m_systemd_notify_interval . do_call ( [ this ] { sd_notify ( 0 , ( " WATCHDOG=1 \n STATUS= " + get_status_string ( ) ) . c_str ( ) ) ; } ) ;
2020-01-19 23:44:18 +01:00
# endif
2014-03-03 23:07:58 +01:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2017-11-12 00:25:12 +01:00
bool core : : check_disk_space ( )
{
uint64_t free_space = get_free_space ( ) ;
if ( free_space < 1ull * 1024 * 1024 * 1024 ) // 1 GB
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , fmt : : format ( fg ( fmt : : terminal_color : : red ) , " Free space is below 1 GB on {} " , m_config_folder ) ) ;
2017-11-12 00:25:12 +01:00
return true ;
}
//-----------------------------------------------------------------------------------------------
2018-09-20 15:31:45 +02:00
double factorial ( unsigned int n )
{
if ( n < = 1 )
return 1.0 ;
double f = n ;
while ( n - - > 1 )
f * = n ;
return f ;
}
//-----------------------------------------------------------------------------------------------
2019-03-19 17:30:17 +01:00
static double probability1 ( unsigned int blocks , unsigned int expected )
2018-09-20 15:31:45 +02:00
{
// https://www.umass.edu/wsp/resources/poisson/#computing
return pow ( expected , blocks ) / ( factorial ( blocks ) * exp ( expected ) ) ;
}
//-----------------------------------------------------------------------------------------------
2019-03-19 17:30:17 +01:00
static double probability ( unsigned int blocks , unsigned int expected )
{
double p = 0.0 ;
if ( blocks < = expected )
{
for ( unsigned int b = 0 ; b < = blocks ; + + b )
p + = probability1 ( b , expected ) ;
}
else if ( blocks > expected )
{
for ( unsigned int b = blocks ; b < = expected * 3 /* close enough */ ; + + b )
p + = probability1 ( b , expected ) ;
}
return p ;
}
//-----------------------------------------------------------------------------------------------
2018-09-20 15:31:45 +02:00
bool core : : check_block_rate ( )
{
2022-05-16 22:55:05 +02:00
if ( m_offline | | m_nettype = = network_type : : FAKECHAIN | | m_target_blockchain_height > get_current_blockchain_height ( ) | | m_target_blockchain_height = = 0 )
2018-09-20 15:31:45 +02:00
{
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " Not checking block rate, offline or syncing " ) ;
2018-09-20 15:31:45 +02:00
return true ;
}
2020-08-07 01:44:23 +02:00
static constexpr double threshold = 1. / ( ( 24 h * 10 ) / TARGET_BLOCK_TIME ) ; // one false positive every 10 days
2019-10-28 17:04:43 +01:00
static constexpr unsigned int max_blocks_checked = 150 ;
2018-09-20 15:31:45 +02:00
const time_t now = time ( NULL ) ;
2019-10-28 17:04:43 +01:00
const std : : vector < time_t > timestamps = m_blockchain_storage . get_last_block_timestamps ( max_blocks_checked ) ;
2018-09-20 15:31:45 +02:00
2019-02-06 20:04:47 +01:00
static const unsigned int seconds [ ] = { 5400 , 3600 , 1800 , 1200 , 600 } ;
2018-09-20 15:31:45 +02:00
for ( size_t n = 0 ; n < sizeof ( seconds ) / sizeof ( seconds [ 0 ] ) ; + + n )
{
unsigned int b = 0 ;
2018-11-05 01:13:51 +01:00
const time_t time_boundary = now - static_cast < time_t > ( seconds [ n ] ) ;
for ( time_t ts : timestamps ) b + = ts > = time_boundary ;
2020-08-07 01:44:23 +02:00
const double p = probability ( b , seconds [ n ] / tools : : to_seconds ( TARGET_BLOCK_TIME ) ) ;
2022-09-13 08:57:56 +02:00
oxen : : log : : debug ( logcat , " blocks in the last {} minutes: {} (probability {}) " , seconds [ n ] / 60 , b , p ) ;
2018-09-20 15:31:45 +02:00
if ( p < threshold )
{
2022-09-13 08:57:56 +02:00
oxen : : log : : warning ( logcat , " There were {}{} blocks in the last {} minutes, \
there might be large hash rate changes , or we might be partitioned , \
cut off from the Loki network or under attack , or your computer ' s time is off . \
Or it could be just sheer bad luck . " , b, (b == max_blocks_checked ? " or more " : " " ), seconds[n] / 60);
2018-09-20 15:31:45 +02:00
break ; // no need to look further
}
}
return true ;
}
//-----------------------------------------------------------------------------------------------
2019-10-22 17:28:08 +02:00
void core : : flush_bad_txs_cache ( )
{
bad_semantics_txes_lock . lock ( ) ;
for ( int idx = 0 ; idx < 2 ; + + idx )
bad_semantics_txes [ idx ] . clear ( ) ;
bad_semantics_txes_lock . unlock ( ) ;
}
//-----------------------------------------------------------------------------------------------
2019-12-02 19:51:28 +01:00
void core : : flush_invalid_blocks ( )
{
m_blockchain_storage . flush_invalid_blocks ( ) ;
}
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
bool core : : update_blockchain_pruning ( )
{
return m_blockchain_storage . update_blockchain_pruning ( ) ;
}
//-----------------------------------------------------------------------------------------------
bool core : : check_blockchain_pruning ( )
{
return m_blockchain_storage . check_blockchain_pruning ( ) ;
}
//-----------------------------------------------------------------------------------------------
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
void core : : set_target_blockchain_height ( uint64_t target_blockchain_height )
{
2016-09-25 20:08:31 +02:00
m_target_blockchain_height = target_blockchain_height ;
2014-06-04 22:50:13 +02:00
}
//-----------------------------------------------------------------------------------------------
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY)
Bockchain:
1. Optim: Multi-thread long-hash computation when encountering groups of blocks.
2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible.
3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible.
4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks.
5. Optim: Multi-thread signature computation whenever possible.
6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD)
7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???).
8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads).
Berkeley-DB:
1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc).
2. Fix: Unable to pop blocks on reorganize due to transaction errors.
3. Patch: Large number of transaction aborts when running multi-threaded bulk queries.
4. Patch: Insufficient locks error when running full sync.
5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation.
6. Optim: Add bulk queries to get output global indices.
7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
9. Optim: Added thread-safe buffers used when multi-threading bulk queries.
10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details)
11. Mod: Added checkpoint thread and auto-remove-logs option.
12. *Now usable on 32-bit systems like RPI2.
LMDB:
1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect)
2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3)
3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key
4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details)
5. Mod: Auto resize to +1GB instead of multiplier x1.5
ETC:
1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete.
2. Fix: 32-bit saturation bug when computing next difficulty on large blocks.
[PENDING ISSUES]
1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization.
This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD.
2. Berkeley db, possible bug "unable to allocate memory". TBD.
[NEW OPTIONS] (*Currently all enabled for testing purposes)
1. --fast-block-sync arg=[0:1] (default: 1)
a. 0 = Compute long hash per block (may take a while depending on CPU)
b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence)
2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000)
a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions.
b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache.
Fast - Write meta-data but defer data flush.
Fastest - Defer meta-data and data flush.
Sync - Flush data after nblocks_per_sync and wait.
Async - Flush data after nblocks_per_sync but do not wait for the operation to finish.
3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower)
Max number of threads to use when computing long-hash in groups.
4. --show-time-stats arg=[0:1] (default: 1)
Show benchmark related time stats.
5. --db-auto-remove-logs arg=[0:1] (default: 1)
For berkeley-db only. Auto remove logs if enabled.
**Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version.
At the moment, you need a full resync to use this optimized version.
[PERFORMANCE COMPARISON]
**Some figures are approximations only.
Using a baseline machine of an i7-2600K+SSD+(with full pow computation):
1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain.
2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain.
3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain.
Averate procesing times (with full pow computation):
lmdb-optimized:
1. tx_ave = 2.5 ms / tx
2. block_ave = 5.87 ms / block
memory-official-repo:
1. tx_ave = 8.85 ms / tx
2. block_ave = 19.68 ms / block
lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e)
1. tx_ave = 47.8 ms / tx
2. block_ave = 64.2 ms / block
**Note: The following data denotes processing times only (does not include p2p download time)
lmdb-optimized processing times (with full pow computation):
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000).
2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000).
3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000).
lmdb-optimized processing times (with per-block-checkpoint)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with full pow computation)
1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000).
2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
berkeley-db optimized processing times (with per-block-checkpoint)
1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 22:09:32 +02:00
uint64_t core : : get_target_blockchain_height ( ) const
{
2014-06-04 22:50:13 +02:00
return m_target_blockchain_height ;
}
2014-09-30 21:24:43 +02:00
//-----------------------------------------------------------------------------------------------
2018-04-16 01:16:02 +02:00
uint64_t core : : prevalidate_block_hashes ( uint64_t height , const std : : vector < crypto : : hash > & hashes )
2017-09-13 22:03:04 +02:00
{
return get_blockchain_storage ( ) . prevalidate_block_hashes ( height , hashes ) ;
}
//-----------------------------------------------------------------------------------------------
2017-11-12 00:25:12 +01:00
uint64_t core : : get_free_space ( ) const
{
2020-10-22 19:55:33 +02:00
return fs : : space ( m_config_folder ) . available ;
2017-11-12 00:25:12 +01:00
}
//-----------------------------------------------------------------------------------------------
2019-10-15 19:52:49 +02:00
std : : shared_ptr < const service_nodes : : quorum > core : : get_quorum ( service_nodes : : quorum_type type , uint64_t height , bool include_old , std : : vector < std : : shared_ptr < const service_nodes : : quorum > > * alt_states ) const
2019-05-01 08:01:17 +02:00
{
2019-10-15 19:52:49 +02:00
return m_service_node_list . get_quorum ( type , height , include_old , alt_states ) ;
2018-10-10 07:24:01 +02:00
}
//-----------------------------------------------------------------------------------------------
Relax deregistration rules
The replaces the deregistration mechanism with a new state change
mechanism (beginning at the v12 fork) which can change a service node's
network status via three potential values (and is extensible in the
future to handle more):
- deregistered -- this is the same as the existing deregistration; the
SN is instantly removed from the SN list.
- decommissioned -- this is a sort of temporary deregistration: your SN
remains in the service node list, but is removed from the rewards list
and from any network duties.
- recommissioned -- this tx is sent by a quorum if they observe a
decommissioned SN sending uptime proofs again. Upon reception, the SN
is reactivated and put on the end of the reward list.
Since this is broadening the quorum use, this also renames the relevant
quorum to a "obligations" quorum (since it validates SN obligations),
while the transactions are "state_change" transactions (since they
change the state of a registered SN).
The new parameters added to service_node_rules.h control how this works:
// Service node decommissioning: as service nodes stay up they earn "credits" (measured in blocks)
// towards a future outage. A new service node starts out with INITIAL_CREDIT, and then builds up
// CREDIT_PER_DAY for each day the service node remains active up to a maximum of
// DECOMMISSION_MAX_CREDIT.
//
// If a service node stops sending uptime proofs, a quorum will consider whether the service node
// has built up enough credits (at least MINIMUM): if so, instead of submitting a deregistration,
// it instead submits a decommission. This removes the service node from the list of active
// service nodes both for rewards and for any active network duties. If the service node comes
// back online (i.e. starts sending the required performance proofs again) before the credits run
// out then a quorum will reinstate the service node using a recommission transaction, which adds
// the service node back to the bottom of the service node reward list, and resets its accumulated
// credits to 0. If it does not come back online within the required number of blocks (i.e. the
// accumulated credit at the point of decommissioning) then a quorum will send a permanent
// deregistration transaction to the network, starting a 30-day deregistration count down.
This commit currently includes values (which are not necessarily
finalized):
- 8 hours (240 blocks) of credit required for activation of a
decommission (rather than a deregister)
- 0 initial credits at registration
- a maximum of 24 hours (720 blocks) of credits
- credits accumulate at a rate that you hit 24 hours of credits after 30
days of operation.
Miscellaneous other details of this PR:
- a new TX extra tag is used for the state change (including
deregistrations). The old extra tag has no version or type tag, so
couldn't be reused. The data in the new tag is slightly more
efficiently packed than the old deregistration transaction, so it gets
used for deregistrations (starting at the v12 fork) as well.
- Correct validator/worker selection required generalizing the shuffle
function to be able to shuffle just part of a vector. This lets us
stick any down service nodes at the end of the potential list, then
select validators by only shuffling the part of the index vector that
contains active service indices. Once the validators are selected, the
remainder of the list (this time including decommissioned SN indices) is
shuffled to select quorum workers to check, thus allowing decommisioned
nodes to be randomly included in the nodes to check without being
selected as a validator.
- Swarm recalculation was not quite right: swarms were recalculated on
SN registrations, even if those registrations were include shared node
registrations, but *not* recalculated on stakes. Starting with the
upgrade this behaviour is fixed (swarms aren't actually used currently
and aren't consensus-relevant so recalculating early won't hurt
anything).
- Details on decomm/dereg are added to RPC info and print_sn/print_sn_status
- Slightly improves the % of reward output in the print_sn output by
rounding it to two digits, and reserves space in the output string to
avoid excessive reallocations.
- Adds various debugging at higher debug levels to quorum voting (into
all of voting itself, vote transmission, and vote reception).
- Reset service node list internal data structure version to 0. The SN
list has to be rescanned anyway at upgrade (its size has changed), so we
might as well reset the version and remove the version-dependent
serialization code. (Note that the affected code here is for SN states
in lmdb storage, not for SN-to-SN communication serialization).
2019-06-18 23:57:02 +02:00
bool core : : is_service_node ( const crypto : : public_key & pubkey , bool require_active ) const
2018-10-10 07:24:01 +02:00
{
Relax deregistration rules
The replaces the deregistration mechanism with a new state change
mechanism (beginning at the v12 fork) which can change a service node's
network status via three potential values (and is extensible in the
future to handle more):
- deregistered -- this is the same as the existing deregistration; the
SN is instantly removed from the SN list.
- decommissioned -- this is a sort of temporary deregistration: your SN
remains in the service node list, but is removed from the rewards list
and from any network duties.
- recommissioned -- this tx is sent by a quorum if they observe a
decommissioned SN sending uptime proofs again. Upon reception, the SN
is reactivated and put on the end of the reward list.
Since this is broadening the quorum use, this also renames the relevant
quorum to a "obligations" quorum (since it validates SN obligations),
while the transactions are "state_change" transactions (since they
change the state of a registered SN).
The new parameters added to service_node_rules.h control how this works:
// Service node decommissioning: as service nodes stay up they earn "credits" (measured in blocks)
// towards a future outage. A new service node starts out with INITIAL_CREDIT, and then builds up
// CREDIT_PER_DAY for each day the service node remains active up to a maximum of
// DECOMMISSION_MAX_CREDIT.
//
// If a service node stops sending uptime proofs, a quorum will consider whether the service node
// has built up enough credits (at least MINIMUM): if so, instead of submitting a deregistration,
// it instead submits a decommission. This removes the service node from the list of active
// service nodes both for rewards and for any active network duties. If the service node comes
// back online (i.e. starts sending the required performance proofs again) before the credits run
// out then a quorum will reinstate the service node using a recommission transaction, which adds
// the service node back to the bottom of the service node reward list, and resets its accumulated
// credits to 0. If it does not come back online within the required number of blocks (i.e. the
// accumulated credit at the point of decommissioning) then a quorum will send a permanent
// deregistration transaction to the network, starting a 30-day deregistration count down.
This commit currently includes values (which are not necessarily
finalized):
- 8 hours (240 blocks) of credit required for activation of a
decommission (rather than a deregister)
- 0 initial credits at registration
- a maximum of 24 hours (720 blocks) of credits
- credits accumulate at a rate that you hit 24 hours of credits after 30
days of operation.
Miscellaneous other details of this PR:
- a new TX extra tag is used for the state change (including
deregistrations). The old extra tag has no version or type tag, so
couldn't be reused. The data in the new tag is slightly more
efficiently packed than the old deregistration transaction, so it gets
used for deregistrations (starting at the v12 fork) as well.
- Correct validator/worker selection required generalizing the shuffle
function to be able to shuffle just part of a vector. This lets us
stick any down service nodes at the end of the potential list, then
select validators by only shuffling the part of the index vector that
contains active service indices. Once the validators are selected, the
remainder of the list (this time including decommissioned SN indices) is
shuffled to select quorum workers to check, thus allowing decommisioned
nodes to be randomly included in the nodes to check without being
selected as a validator.
- Swarm recalculation was not quite right: swarms were recalculated on
SN registrations, even if those registrations were include shared node
registrations, but *not* recalculated on stakes. Starting with the
upgrade this behaviour is fixed (swarms aren't actually used currently
and aren't consensus-relevant so recalculating early won't hurt
anything).
- Details on decomm/dereg are added to RPC info and print_sn/print_sn_status
- Slightly improves the % of reward output in the print_sn output by
rounding it to two digits, and reserves space in the output string to
avoid excessive reallocations.
- Adds various debugging at higher debug levels to quorum voting (into
all of voting itself, vote transmission, and vote reception).
- Reset service node list internal data structure version to 0. The SN
list has to be rescanned anyway at upgrade (its size has changed), so we
might as well reset the version and remove the version-dependent
serialization code. (Note that the affected code here is for SN states
in lmdb storage, not for SN-to-SN communication serialization).
2019-06-18 23:57:02 +02:00
return m_service_node_list . is_service_node ( pubkey , require_active ) ;
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
}
//-----------------------------------------------------------------------------------------------
2019-01-25 04:15:52 +01:00
const std : : vector < service_nodes : : key_image_blacklist_entry > & core : : get_service_node_blacklisted_key_images ( ) const
{
2019-08-11 05:46:59 +02:00
return m_service_node_list . get_blacklisted_key_images ( ) ;
2019-01-25 04:15:52 +01:00
}
//-----------------------------------------------------------------------------------------------
2018-08-15 04:59:05 +02:00
std : : vector < service_nodes : : service_node_pubkey_info > core : : get_service_node_list_state ( const std : : vector < crypto : : public_key > & service_node_pubkeys ) const
{
2019-08-11 05:46:59 +02:00
return m_service_node_list . get_service_node_list_state ( service_node_pubkeys ) ;
2018-08-15 04:59:05 +02:00
}
//-----------------------------------------------------------------------------------------------
2019-05-31 03:06:42 +02:00
bool core : : add_service_node_vote ( const service_nodes : : quorum_vote_t & vote , vote_verification_context & vvc )
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
{
2019-08-11 05:46:59 +02:00
return m_quorum_cop . handle_vote ( vote , vvc ) ;
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
}
2018-08-29 07:53:13 +02:00
//-----------------------------------------------------------------------------------------------
Pruning
The blockchain prunes seven eighths of prunable tx data.
This saves about two thirds of the blockchain size, while
keeping the node useful as a sync source for an eighth
of the blockchain.
No other data is currently pruned.
There are three ways to prune a blockchain:
- run monerod with --prune-blockchain
- run "prune_blockchain" in the monerod console
- run the monero-blockchain-prune utility
The first two will prune in place. Due to how LMDB works, this
will not reduce the blockchain size on disk. Instead, it will
mark parts of the file as free, so that future data will use
that free space, causing the file to not grow until free space
grows scarce.
The third way will create a second database, a pruned copy of
the original one. Since this is a new file, this one will be
smaller than the original one.
Once the database is pruned, it will stay pruned as it syncs.
That is, there is no need to use --prune-blockchain again, etc.
2018-04-30 00:30:51 +02:00
uint32_t core : : get_blockchain_pruning_seed ( ) const
{
return get_blockchain_storage ( ) . get_blockchain_pruning_seed ( ) ;
}
//-----------------------------------------------------------------------------------------------
bool core : : prune_blockchain ( uint32_t pruning_seed )
{
return get_blockchain_storage ( ) . prune_blockchain ( pruning_seed ) ;
}
Service Node Deregister Part 5 (#89)
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* core, service_node_list: separated address from service node pubkey
* Retrieve quorum list from height, reviewed
* Setup data structures for de/register TX
* Submit and validate partial/full deregisters
* Add P2P relaying of partial deregistration votes
* Code review adjustments for deregistration part 1
- Fix check_tx_semantic
- Remove signature_pod as votes are now stored as blobs. Serialization
overrides don't intefere with crypto::signature anymore.
* deregistration_vote_pool - changed sign/verify interface and removed repeated code
* Misc review, fix sign/verify api, vote threshold
* Deregister/tx edge case handling for combinatoric votes
* Store service node lists for the duration of deregister lifetimes
* Quorum min/max bug, sort node list, fix node to test list
* Change quorum to store acc pub address, fix oob bug
* Code review for expiring votes, acc keys to pub_key, improve err msgs
* Add early out for is_deregistration_tx and protect against quorum changes
* Remove debug code, fix segfault
* Remove irrelevant check for tx v3 in blockchain, fix >= height for pruning quorum states
Incorrect assumption that a transaction can be kept in the chain if it could
eventually become invalid, because if it were the chain would be split and
eventually these transaction would be dropped. But also that we should not
override the pre-existing logic which handles this case anyway.
2018-07-18 04:42:47 +02:00
//-----------------------------------------------------------------------------------------------
2017-01-09 00:50:29 +01:00
std : : time_t core : : get_start_time ( ) const
{
return start_time ;
}
//-----------------------------------------------------------------------------------------------
2014-09-30 21:24:43 +02:00
void core : : graceful_exit ( )
{
raise ( SIGTERM ) ;
}
2014-03-03 23:07:58 +01:00
}