2014-07-25 16:29:08 +00:00
// Copyright (c) 2006-2013, Andrey N. Sabelnikov, www.sabelnikov.net
2014-03-03 22:07:58 +00:00
// All rights reserved.
//
2014-07-25 16:29:08 +00:00
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * 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.
// * Neither the name of the Andrey N. Sabelnikov nor the
// names of its contributors may be used to endorse or promote products
// derived from this software without specific prior written permission.
2014-03-03 22:07:58 +00:00
//
2014-07-25 16:29:08 +00:00
// 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 OWNER 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.
2014-03-03 22:07:58 +00:00
//
# pragma once
2017-03-19 02:05:14 +00:00
# include <boost/asio/deadline_timer.hpp>
2014-03-03 22:07:58 +00:00
# include <boost/uuid/uuid_generators.hpp>
2016-01-27 14:07:14 +00:00
# include <boost/unordered_map.hpp>
2014-03-03 22:07:58 +00:00
# include <boost/interprocess/detail/atomic.hpp>
# include <boost/smart_ptr/make_shared.hpp>
# include <atomic>
2019-05-16 20:34:22 +00:00
# include <deque>
2014-03-03 22:07:58 +00:00
# include "levin_base.h"
2018-12-06 18:04:33 +00:00
# include "buffer.h"
2014-03-03 22:07:58 +00:00
# include "misc_language.h"
2017-11-25 22:25:05 +00:00
# include "syncobj.h"
# include "misc_os_dependent.h"
2018-11-18 10:18:35 +00:00
# include "int-util.h"
2014-03-03 22:07:58 +00:00
2015-04-01 18:25:53 +00:00
# include <random>
# include <chrono>
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 16:34:23 +00:00
# undef MONERO_DEFAULT_LOG_CATEGORY
# define MONERO_DEFAULT_LOG_CATEGORY "net"
2014-03-03 22:07:58 +00:00
2017-06-04 21:36:09 +00:00
# ifndef MIN_BYTES_WANTED
# define MIN_BYTES_WANTED 512
# endif
2021-01-19 02:22:32 +00:00
template < typename context_t >
void on_levin_traffic ( const context_t & context , bool initiator , bool sent , bool error , size_t bytes , const char * category )
{
MCINFO ( " net.p2p.traffic " , context < < bytes < < " bytes " < < ( sent ? " sent " : " received " ) < < ( error ? " /corrupt " : " " )
< < " for category " < < category < < " initiated by " < < ( initiator ? " us " : " peer " ) ) ;
}
template < typename context_t >
void on_levin_traffic ( const context_t & context , bool initiator , bool sent , bool error , size_t bytes , int command )
{
char buf [ 32 ] ;
snprintf ( buf , sizeof ( buf ) , " command-%u " , command ) ;
on_levin_traffic ( context , initiator , sent , error , bytes , buf ) ;
}
2014-03-03 22:07:58 +00:00
namespace epee
{
namespace levin
{
/************************************************************************/
/* */
/************************************************************************/
template < class t_connection_context >
class async_protocol_handler ;
template < class t_connection_context >
class async_protocol_handler_config
{
2016-01-27 14:07:14 +00:00
typedef boost : : unordered_map < boost : : uuids : : uuid , async_protocol_handler < t_connection_context > * > connections_map ;
2014-03-03 22:07:58 +00:00
critical_section m_connects_lock ;
connections_map m_connects ;
void add_connection ( async_protocol_handler < t_connection_context > * pc ) ;
void del_connection ( async_protocol_handler < t_connection_context > * pc ) ;
async_protocol_handler < t_connection_context > * find_connection ( boost : : uuids : : uuid connection_id ) const ;
int find_and_lock_connection ( boost : : uuids : : uuid connection_id , async_protocol_handler < t_connection_context > * & aph ) ;
friend class async_protocol_handler < t_connection_context > ;
2017-10-09 15:46:42 +00:00
levin_commands_handler < t_connection_context > * m_pcommands_handler ;
void ( * m_pcommands_handler_destroy ) ( levin_commands_handler < t_connection_context > * ) ;
2018-01-20 21:44:23 +00:00
void delete_connections ( size_t count , bool incoming ) ;
2014-03-03 22:07:58 +00:00
public :
typedef t_connection_context connection_context ;
2020-12-30 00:58:53 +00:00
uint64_t m_initial_max_packet_size ;
uint64_t m_max_packet_size ;
2014-03-20 11:46:11 +00:00
uint64_t m_invoke_timeout ;
2014-03-03 22:07:58 +00:00
2021-01-19 02:22:32 +00:00
int invoke ( int command , message_writer in_msg , std : : string & buff_out , boost : : uuids : : uuid connection_id ) ;
2014-03-03 22:07:58 +00:00
template < class callback_t >
2021-01-19 02:22:32 +00:00
int invoke_async ( int command , message_writer in_msg , boost : : uuids : : uuid connection_id , const callback_t & cb , size_t timeout = LEVIN_DEFAULT_TIMEOUT_PRECONFIGURED ) ;
2014-03-03 22:07:58 +00:00
2019-05-16 20:34:22 +00:00
int send ( epee : : byte_slice message , const boost : : uuids : : uuid & connection_id ) ;
2014-03-03 22:07:58 +00:00
bool close ( boost : : uuids : : uuid connection_id ) ;
bool update_connection_context ( const t_connection_context & contxt ) ;
bool request_callback ( boost : : uuids : : uuid connection_id ) ;
template < class callback_t >
2018-02-02 18:45:12 +00:00
bool foreach_connection ( const callback_t & cb ) ;
2017-07-02 21:41:15 +00:00
template < class callback_t >
2018-02-02 18:45:12 +00:00
bool for_connection ( const boost : : uuids : : uuid & connection_id , const callback_t & cb ) ;
2014-03-03 22:07:58 +00:00
size_t get_connections_count ( ) ;
2019-06-18 22:47:05 +00:00
size_t get_out_connections_count ( ) ;
size_t get_in_connections_count ( ) ;
2017-10-09 15:46:42 +00:00
void set_handler ( levin_commands_handler < t_connection_context > * handler , void ( * destroy ) ( levin_commands_handler < t_connection_context > * ) = NULL ) ;
2014-03-03 22:07:58 +00:00
2020-12-30 00:58:53 +00:00
async_protocol_handler_config ( ) : m_pcommands_handler ( NULL ) , m_pcommands_handler_destroy ( NULL ) , m_initial_max_packet_size ( LEVIN_INITIAL_MAX_PACKET_SIZE ) , m_max_packet_size ( LEVIN_DEFAULT_MAX_PACKET_SIZE ) , m_invoke_timeout ( LEVIN_DEFAULT_TIMEOUT_PRECONFIGURED )
2014-03-03 22:07:58 +00:00
{ }
2017-12-11 18:33:37 +00:00
~ async_protocol_handler_config ( ) { set_handler ( NULL , NULL ) ; }
2015-02-12 19:59:39 +00:00
void del_out_connections ( size_t count ) ;
2018-01-20 21:44:23 +00:00
void del_in_connections ( size_t count ) ;
2014-03-03 22:07:58 +00:00
} ;
/************************************************************************/
/* */
/************************************************************************/
template < class t_connection_context = net_utils : : connection_context_base >
class async_protocol_handler
{
2019-05-16 20:34:22 +00:00
std : : string m_fragment_buffer ;
2021-01-19 02:22:32 +00:00
bool send_message ( byte_slice message )
2019-05-16 20:34:22 +00:00
{
2021-01-19 02:22:32 +00:00
if ( message . size ( ) < sizeof ( message_writer : : header ) )
return false ;
message_writer : : header head ;
std : : memcpy ( std : : addressof ( head ) , message . data ( ) , sizeof ( head ) ) ;
if ( ! m_pservice_endpoint - > do_send ( std : : move ( message ) ) )
2019-05-16 20:34:22 +00:00
return false ;
2021-01-19 02:22:32 +00:00
on_levin_traffic ( m_connection_context , true , true , false , head . m_cb , head . m_command ) ;
2019-05-16 20:34:22 +00:00
MDEBUG ( m_connection_context < < " LEVIN_PACKET_SENT. [len= " < < head . m_cb
< < " , flags " < < head . m_flags
< < " , r?= " < < head . m_have_to_return_data
< < " , cmd = " < < head . m_command
< < " , ver= " < < head . m_protocol_version ) ;
return true ;
}
2014-03-03 22:07:58 +00:00
public :
typedef t_connection_context connection_context ;
typedef async_protocol_handler_config < t_connection_context > config_type ;
enum stream_state
{
stream_state_head ,
stream_state_body
} ;
std : : atomic < bool > m_protocol_released ;
volatile uint32_t m_invoke_buf_ready ;
volatile int m_invoke_result_code ;
critical_section m_local_inv_buff_lock ;
std : : string m_local_inv_buff ;
critical_section m_call_lock ;
volatile uint32_t m_wait_count ;
volatile uint32_t m_close_called ;
bucket_head2 m_current_head ;
net_utils : : i_service_endpoint * m_pservice_endpoint ;
config_type & m_config ;
t_connection_context & m_connection_context ;
2020-12-30 00:58:53 +00:00
std : : atomic < uint64_t > m_max_packet_size ;
2014-03-03 22:07:58 +00:00
2018-12-06 18:04:33 +00:00
net_utils : : buffer m_cache_in_buffer ;
2014-03-03 22:07:58 +00:00
stream_state m_state ;
2014-03-20 11:46:11 +00:00
int32_t m_oponent_protocol_ver ;
2014-03-03 22:07:58 +00:00
bool m_connection_initialized ;
struct invoke_response_handler_base
{
2018-12-06 18:04:33 +00:00
virtual bool handle ( int res , const epee : : span < const uint8_t > buff , connection_context & context ) = 0 ;
2014-03-03 22:07:58 +00:00
virtual bool is_timer_started ( ) const = 0 ;
virtual void cancel ( ) = 0 ;
virtual bool cancel_timer ( ) = 0 ;
2017-06-04 21:36:09 +00:00
virtual void reset_timer ( ) = 0 ;
2014-03-03 22:07:58 +00:00
} ;
template < class callback_t >
struct anvoke_handler : invoke_response_handler_base
{
anvoke_handler ( const callback_t & cb , uint64_t timeout , async_protocol_handler & con , int command )
2017-06-04 21:36:09 +00:00
: m_cb ( cb ) , m_timeout ( timeout ) , m_con ( con ) , m_timer ( con . m_pservice_endpoint - > get_io_service ( ) ) , m_timer_started ( false ) ,
2014-03-03 22:07:58 +00:00
m_cancel_timer_called ( false ) , m_timer_cancelled ( false ) , m_command ( command )
{
if ( m_con . start_outer_call ( ) )
{
2017-06-04 21:36:09 +00:00
MDEBUG ( con . get_context_ref ( ) < < " anvoke_handler, timeout: " < < timeout ) ;
2014-03-03 22:07:58 +00:00
m_timer . expires_from_now ( boost : : posix_time : : milliseconds ( timeout ) ) ;
2017-06-28 10:24:34 +00:00
m_timer . async_wait ( [ & con , command , cb , timeout ] ( const boost : : system : : error_code & ec )
2014-03-03 22:07:58 +00:00
{
2017-06-28 10:24:34 +00:00
if ( ec = = boost : : asio : : error : : operation_aborted )
return ;
MINFO ( con . get_context_ref ( ) < < " Timeout on invoke operation happened, command: " < < command < < " timeout: " < < timeout ) ;
2018-12-06 18:04:33 +00:00
epee : : span < const uint8_t > fake ;
2017-06-28 10:24:34 +00:00
cb ( LEVIN_ERROR_CONNECTION_TIMEDOUT , fake , con . get_context_ref ( ) ) ;
con . close ( ) ;
con . finish_outer_call ( ) ;
2014-03-03 22:07:58 +00:00
} ) ;
m_timer_started = true ;
}
}
virtual ~ anvoke_handler ( )
{ }
callback_t m_cb ;
async_protocol_handler & m_con ;
boost : : asio : : deadline_timer m_timer ;
bool m_timer_started ;
bool m_cancel_timer_called ;
bool m_timer_cancelled ;
2017-06-04 21:36:09 +00:00
uint64_t m_timeout ;
2014-03-03 22:07:58 +00:00
int m_command ;
2018-12-06 18:04:33 +00:00
virtual bool handle ( int res , const epee : : span < const uint8_t > buff , typename async_protocol_handler : : connection_context & context )
2014-03-03 22:07:58 +00:00
{
if ( ! cancel_timer ( ) )
return false ;
m_cb ( res , buff , context ) ;
m_con . finish_outer_call ( ) ;
return true ;
}
virtual bool is_timer_started ( ) const
{
return m_timer_started ;
}
virtual void cancel ( )
{
if ( cancel_timer ( ) )
{
2018-12-06 18:04:33 +00:00
epee : : span < const uint8_t > fake ;
2014-03-03 22:07:58 +00:00
m_cb ( LEVIN_ERROR_CONNECTION_DESTROYED , fake , m_con . get_context_ref ( ) ) ;
m_con . finish_outer_call ( ) ;
}
}
virtual bool cancel_timer ( )
{
if ( ! m_cancel_timer_called )
{
m_cancel_timer_called = true ;
boost : : system : : error_code ignored_ec ;
m_timer_cancelled = 1 = = m_timer . cancel ( ignored_ec ) ;
}
return m_timer_cancelled ;
}
2017-06-04 21:36:09 +00:00
virtual void reset_timer ( )
{
boost : : system : : error_code ignored_ec ;
if ( ! m_cancel_timer_called & & m_timer . cancel ( ignored_ec ) > 0 )
{
2017-06-28 10:24:34 +00:00
callback_t & cb = m_cb ;
uint64_t timeout = m_timeout ;
async_protocol_handler & con = m_con ;
int command = m_command ;
2017-06-04 21:36:09 +00:00
m_timer . expires_from_now ( boost : : posix_time : : milliseconds ( m_timeout ) ) ;
2017-06-28 10:24:34 +00:00
m_timer . async_wait ( [ & con , cb , command , timeout ] ( const boost : : system : : error_code & ec )
2017-06-04 21:36:09 +00:00
{
2017-06-28 10:24:34 +00:00
if ( ec = = boost : : asio : : error : : operation_aborted )
return ;
MINFO ( con . get_context_ref ( ) < < " Timeout on invoke operation happened, command: " < < command < < " timeout: " < < timeout ) ;
2018-12-06 18:04:33 +00:00
epee : : span < const uint8_t > fake ;
2017-06-28 10:24:34 +00:00
cb ( LEVIN_ERROR_CONNECTION_TIMEDOUT , fake , con . get_context_ref ( ) ) ;
con . close ( ) ;
con . finish_outer_call ( ) ;
2017-06-04 21:36:09 +00:00
} ) ;
}
}
2014-03-03 22:07:58 +00:00
} ;
critical_section m_invoke_response_handlers_lock ;
std : : list < boost : : shared_ptr < invoke_response_handler_base > > m_invoke_response_handlers ;
template < class callback_t >
2018-02-02 18:45:12 +00:00
bool add_invoke_response_handler ( const callback_t & cb , uint64_t timeout , async_protocol_handler & con , int command )
2014-03-03 22:07:58 +00:00
{
CRITICAL_REGION_LOCAL ( m_invoke_response_handlers_lock ) ;
2019-09-12 12:29:47 +00:00
if ( m_protocol_released )
{
MERROR ( " Adding response handler to a released object " ) ;
return false ;
}
2014-03-03 22:07:58 +00:00
boost : : shared_ptr < invoke_response_handler_base > handler ( boost : : make_shared < anvoke_handler < callback_t > > ( cb , timeout , con , command ) ) ;
m_invoke_response_handlers . push_back ( handler ) ;
return handler - > is_timer_started ( ) ;
}
template < class callback_t > friend struct anvoke_handler ;
public :
async_protocol_handler ( net_utils : : i_service_endpoint * psnd_hndlr ,
config_type & config ,
t_connection_context & conn_context ) :
m_current_head ( bucket_head2 ( ) ) ,
m_pservice_endpoint ( psnd_hndlr ) ,
m_config ( config ) ,
2020-12-30 00:58:53 +00:00
m_connection_context ( conn_context ) ,
m_max_packet_size ( config . m_initial_max_packet_size ) ,
2019-04-22 22:29:56 +00:00
m_cache_in_buffer ( 4 * 1024 ) ,
2014-03-03 22:07:58 +00:00
m_state ( stream_state_head )
{
m_close_called = 0 ;
m_protocol_released = false ;
m_wait_count = 0 ;
m_oponent_protocol_ver = 0 ;
m_connection_initialized = false ;
2018-10-02 12:14:44 +00:00
m_invoke_buf_ready = 0 ;
m_invoke_result_code = LEVIN_ERROR_CONNECTION ;
2014-03-03 22:07:58 +00:00
}
virtual ~ async_protocol_handler ( )
{
2018-09-29 20:17:00 +00:00
try
{
2014-03-03 22:07:58 +00:00
if ( m_connection_initialized )
{
m_config . del_connection ( this ) ;
}
for ( size_t i = 0 ; i < 60 * 1000 / 100 & & 0 ! = boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_wait_count ) ; + + i )
{
misc_utils : : sleep_no_w ( 100 ) ;
}
CHECK_AND_ASSERT_MES_NO_RET ( 0 = = boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_wait_count ) , " Failed to wait for operation completion. m_wait_count = " < < m_wait_count ) ;
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 16:34:23 +00:00
MTRACE ( m_connection_context < < " ~async_protocol_handler() " ) ;
2018-09-29 20:17:00 +00:00
}
catch ( . . . ) { /* ignore */ }
2014-03-03 22:07:58 +00:00
}
bool start_outer_call ( )
{
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 16:34:23 +00:00
MTRACE ( m_connection_context < < " [levin_protocol] -->> start_outer_call " ) ;
2014-03-03 22:07:58 +00:00
if ( ! m_pservice_endpoint - > add_ref ( ) )
{
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 16:34:23 +00:00
MERROR ( m_connection_context < < " [levin_protocol] -->> start_outer_call failed " ) ;
2014-03-03 22:07:58 +00:00
return false ;
}
boost : : interprocess : : ipcdetail : : atomic_inc32 ( & m_wait_count ) ;
return true ;
}
bool finish_outer_call ( )
{
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 16:34:23 +00:00
MTRACE ( m_connection_context < < " [levin_protocol] <<-- finish_outer_call " ) ;
2014-03-03 22:07:58 +00:00
boost : : interprocess : : ipcdetail : : atomic_dec32 ( & m_wait_count ) ;
m_pservice_endpoint - > release ( ) ;
return true ;
}
bool release_protocol ( )
{
decltype ( m_invoke_response_handlers ) local_invoke_response_handlers ;
CRITICAL_REGION_BEGIN ( m_invoke_response_handlers_lock ) ;
local_invoke_response_handlers . swap ( m_invoke_response_handlers ) ;
m_protocol_released = true ;
CRITICAL_REGION_END ( ) ;
// Never call callback inside critical section, that can cause deadlock. Callback can be called when
// invoke_response_handler_base is cancelled
std : : for_each ( local_invoke_response_handlers . begin ( ) , local_invoke_response_handlers . end ( ) , [ ] ( const boost : : shared_ptr < invoke_response_handler_base > & pinv_resp_hndlr ) {
pinv_resp_hndlr - > cancel ( ) ;
} ) ;
return true ;
}
bool close ( )
{
boost : : interprocess : : ipcdetail : : atomic_inc32 ( & m_close_called ) ;
m_pservice_endpoint - > close ( ) ;
return true ;
}
void update_connection_context ( const connection_context & contxt )
{
m_connection_context = contxt ;
}
void request_callback ( )
{
misc_utils : : auto_scope_leave_caller scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler : : finish_outer_call , this ) ) ;
m_pservice_endpoint - > request_callback ( ) ;
}
void handle_qued_callback ( )
{
m_config . m_pcommands_handler - > callback ( m_connection_context ) ;
}
virtual bool handle_recv ( const void * ptr , size_t cb )
{
if ( boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_close_called ) )
return false ; //closing connections
if ( ! m_config . m_pcommands_handler )
{
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 16:34:23 +00:00
MERROR ( m_connection_context < < " Commands handler not set! " ) ;
2014-03-03 22:07:58 +00:00
return false ;
}
2019-05-16 20:34:22 +00:00
// these should never fail, but do runtime check for safety
2020-12-30 00:58:53 +00:00
const uint64_t max_packet_size = m_max_packet_size ;
CHECK_AND_ASSERT_MES ( max_packet_size > = m_cache_in_buffer . size ( ) , false , " Bad m_cache_in_buffer.size() " ) ;
CHECK_AND_ASSERT_MES ( max_packet_size - m_cache_in_buffer . size ( ) > = m_fragment_buffer . size ( ) , false , " Bad m_cache_in_buffer.size() + m_fragment_buffer.size() " ) ;
2019-05-16 20:34:22 +00:00
// flipped to subtraction; prevent overflow since m_max_packet_size is variable and public
2020-12-30 00:58:53 +00:00
if ( cb > max_packet_size - m_cache_in_buffer . size ( ) - m_fragment_buffer . size ( ) )
2014-03-03 22:07:58 +00:00
{
2020-12-30 00:58:53 +00:00
MWARNING ( m_connection_context < < " Maximum packet size exceed!, m_max_packet_size = " < < max_packet_size
2014-03-03 22:07:58 +00:00
< < " , packet received " < < m_cache_in_buffer . size ( ) + cb
< < " , connection will be closed. " ) ;
return false ;
}
m_cache_in_buffer . append ( ( const char * ) ptr , cb ) ;
bool is_continue = true ;
while ( is_continue )
{
switch ( m_state )
{
case stream_state_body :
if ( m_cache_in_buffer . size ( ) < m_current_head . m_cb )
{
is_continue = false ;
2017-12-15 10:27:44 +00:00
if ( cb > = MIN_BYTES_WANTED )
2017-06-04 21:36:09 +00:00
{
2017-12-15 10:27:44 +00:00
CRITICAL_REGION_LOCAL ( m_invoke_response_handlers_lock ) ;
if ( ! m_invoke_response_handlers . empty ( ) )
{
//async call scenario
boost : : shared_ptr < invoke_response_handler_base > response_handler = m_invoke_response_handlers . front ( ) ;
response_handler - > reset_timer ( ) ;
2020-12-31 16:51:59 +00:00
MDEBUG ( m_connection_context < < " LEVIN_PACKET partial msg received. len= " < < cb < < " , current total " < < m_cache_in_buffer . size ( ) < < " / " < < m_current_head . m_cb < < " ( " < < ( 100.0f * m_cache_in_buffer . size ( ) / ( m_current_head . m_cb ? m_current_head . m_cb : 1 ) ) < < " %) " ) ;
2017-12-15 10:27:44 +00:00
}
2017-06-04 21:36:09 +00:00
}
2014-03-03 22:07:58 +00:00
break ;
}
2019-05-16 20:34:22 +00:00
2014-03-03 22:07:58 +00:00
{
2019-05-16 20:34:22 +00:00
std : : string temp { } ;
2018-12-06 18:04:33 +00:00
epee : : span < const uint8_t > buff_to_invoke = m_cache_in_buffer . carve ( ( std : : string : : size_type ) m_current_head . m_cb ) ;
2019-05-16 20:34:22 +00:00
m_state = stream_state_head ;
// abstract_tcp_server2.h manages max bandwidth for a p2p link
if ( ! ( m_current_head . m_flags & ( LEVIN_PACKET_REQUEST | LEVIN_PACKET_RESPONSE ) ) )
{
// special noise/fragment command
static constexpr const uint32_t both_flags = ( LEVIN_PACKET_BEGIN | LEVIN_PACKET_END ) ;
if ( ( m_current_head . m_flags & both_flags ) = = both_flags )
break ; // noise message, skip to next message
if ( m_current_head . m_flags & LEVIN_PACKET_BEGIN )
m_fragment_buffer . clear ( ) ;
m_fragment_buffer . append ( reinterpret_cast < const char * > ( buff_to_invoke . data ( ) ) , buff_to_invoke . size ( ) ) ;
if ( ! ( m_current_head . m_flags & LEVIN_PACKET_END ) )
break ; // skip to next message
if ( m_fragment_buffer . size ( ) < sizeof ( bucket_head2 ) )
{
MERROR ( m_connection_context < < " Fragmented data too small for levin header " ) ;
return false ;
}
temp = std : : move ( m_fragment_buffer ) ;
m_fragment_buffer . clear ( ) ;
std : : memcpy ( std : : addressof ( m_current_head ) , std : : addressof ( temp [ 0 ] ) , sizeof ( bucket_head2 ) ) ;
2020-12-31 16:51:59 +00:00
const size_t max_bytes = m_connection_context . get_max_bytes ( m_current_head . m_command ) ;
if ( m_current_head . m_cb > std : : min < size_t > ( max_packet_size , max_bytes ) )
{
MERROR ( m_connection_context < < " Maximum packet size exceed!, m_max_packet_size = " < < std : : min < size_t > ( max_packet_size , max_bytes )
< < " , packet header received " < < m_current_head . m_cb < < " , command " < < m_current_head . m_command
< < " , connection will be closed. " ) ;
return false ;
}
2019-05-16 20:34:22 +00:00
buff_to_invoke = { reinterpret_cast < const uint8_t * > ( temp . data ( ) ) + sizeof ( bucket_head2 ) , temp . size ( ) - sizeof ( bucket_head2 ) } ;
}
2014-03-03 22:07:58 +00:00
bool is_response = ( m_oponent_protocol_ver = = LEVIN_PROTOCOL_VER_1 & & m_current_head . m_flags & LEVIN_PACKET_RESPONSE ) ;
2018-12-04 06:27:08 +00:00
MDEBUG ( m_connection_context < < " LEVIN_PACKET_RECEIVED. [len= " < < m_current_head . m_cb
2014-03-03 22:07:58 +00:00
< < " , flags " < < m_current_head . m_flags
< < " , r?= " < < m_current_head . m_have_to_return_data
< < " , cmd = " < < m_current_head . m_command
< < " , v= " < < m_current_head . m_protocol_version ) ;
if ( is_response )
{ //response to some invoke
epee : : critical_region_t < decltype ( m_invoke_response_handlers_lock ) > invoke_response_handlers_guard ( m_invoke_response_handlers_lock ) ;
if ( ! m_invoke_response_handlers . empty ( ) )
{ //async call scenario
boost : : shared_ptr < invoke_response_handler_base > response_handler = m_invoke_response_handlers . front ( ) ;
bool timer_cancelled = response_handler - > cancel_timer ( ) ;
// Don't pop handler, to avoid destroying it
if ( timer_cancelled )
m_invoke_response_handlers . pop_front ( ) ;
invoke_response_handlers_guard . unlock ( ) ;
if ( timer_cancelled )
2016-11-12 10:12:23 +00:00
response_handler - > handle ( m_current_head . m_return_code , buff_to_invoke , m_connection_context ) ;
2014-03-03 22:07:58 +00:00
}
else
{
invoke_response_handlers_guard . unlock ( ) ;
//use sync call scenario
if ( ! boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_wait_count ) & & ! boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_close_called ) )
{
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 16:34:23 +00:00
MERROR ( m_connection_context < < " no active invoke when response came, wtf? " ) ;
2014-03-03 22:07:58 +00:00
return false ;
} else
{
CRITICAL_REGION_BEGIN ( m_local_inv_buff_lock ) ;
2018-12-06 18:04:33 +00:00
m_local_inv_buff = std : : string ( ( const char * ) buff_to_invoke . data ( ) , buff_to_invoke . size ( ) ) ;
buff_to_invoke = epee : : span < const uint8_t > ( ( const uint8_t * ) NULL , 0 ) ;
2014-03-03 22:07:58 +00:00
m_invoke_result_code = m_current_head . m_return_code ;
CRITICAL_REGION_END ( ) ;
boost : : interprocess : : ipcdetail : : atomic_write32 ( & m_invoke_buf_ready , 1 ) ;
}
}
} else
{
if ( m_current_head . m_have_to_return_data )
{
2021-01-19 02:22:32 +00:00
levin : : message_writer return_message { 32 * 1024 } ;
2019-05-11 15:38:35 +00:00
const uint32_t return_code = m_config . m_pcommands_handler - > invoke (
2021-01-19 02:22:32 +00:00
m_current_head . m_command , buff_to_invoke , return_message . buffer , m_connection_context
2019-05-11 15:38:35 +00:00
) ;
2020-12-30 00:58:53 +00:00
// peer_id remains unset if dropped
if ( m_current_head . m_command = = m_connection_context . handshake_command ( ) & & m_connection_context . handshake_complete ( ) )
m_max_packet_size = m_config . m_max_packet_size ;
2021-01-19 02:22:32 +00:00
if ( ! send_message ( return_message . finalize_response ( m_current_head . m_command , return_code ) ) )
2014-03-03 22:07:58 +00:00
return false ;
}
else
m_config . m_pcommands_handler - > notify ( m_current_head . m_command , buff_to_invoke , m_connection_context ) ;
}
2019-05-16 20:34:22 +00:00
// reuse small buffer
if ( ! temp . empty ( ) & & temp . capacity ( ) < = 64 * 1024 )
{
temp . clear ( ) ;
m_fragment_buffer = std : : move ( temp ) ;
}
2014-03-03 22:07:58 +00:00
}
break ;
case stream_state_head :
{
if ( m_cache_in_buffer . size ( ) < sizeof ( bucket_head2 ) )
{
2018-12-06 18:04:33 +00:00
if ( m_cache_in_buffer . size ( ) > = sizeof ( uint64_t ) & & * ( ( uint64_t * ) m_cache_in_buffer . span ( 8 ) . data ( ) ) ! = SWAP64LE ( LEVIN_SIGNATURE ) )
2014-03-03 22:07:58 +00:00
{
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 16:34:23 +00:00
MWARNING ( m_connection_context < < " Signature mismatch, connection will be closed " ) ;
2014-03-03 22:07:58 +00:00
return false ;
}
is_continue = false ;
break ;
}
2018-11-18 10:18:35 +00:00
# if BYTE_ORDER == LITTLE_ENDIAN
2018-12-06 18:04:33 +00:00
bucket_head2 & phead = * ( bucket_head2 * ) m_cache_in_buffer . span ( sizeof ( bucket_head2 ) ) . data ( ) ;
2018-11-18 10:18:35 +00:00
# else
2018-12-06 18:04:33 +00:00
bucket_head2 phead = * ( bucket_head2 * ) m_cache_in_buffer . span ( sizeof ( bucket_head2 ) ) . data ( ) ;
2018-11-18 10:18:35 +00:00
phead . m_signature = SWAP64LE ( phead . m_signature ) ;
phead . m_cb = SWAP64LE ( phead . m_cb ) ;
phead . m_command = SWAP32LE ( phead . m_command ) ;
phead . m_return_code = SWAP32LE ( phead . m_return_code ) ;
phead . m_flags = SWAP32LE ( phead . m_flags ) ;
phead . m_protocol_version = SWAP32LE ( phead . m_protocol_version ) ;
# endif
if ( LEVIN_SIGNATURE ! = phead . m_signature )
2014-03-03 22:07:58 +00:00
{
LOG_ERROR_CC ( m_connection_context , " Signature mismatch, connection will be closed " ) ;
return false ;
}
2018-11-18 10:18:35 +00:00
m_current_head = phead ;
2014-03-03 22:07:58 +00:00
2018-12-06 18:04:33 +00:00
m_cache_in_buffer . erase ( sizeof ( bucket_head2 ) ) ;
2014-03-03 22:07:58 +00:00
m_state = stream_state_body ;
m_oponent_protocol_ver = m_current_head . m_protocol_version ;
2020-12-31 16:51:59 +00:00
const size_t max_bytes = m_connection_context . get_max_bytes ( m_current_head . m_command ) ;
if ( m_current_head . m_cb > std : : min < size_t > ( max_packet_size , max_bytes ) )
2014-03-03 22:07:58 +00:00
{
2020-12-31 16:51:59 +00:00
LOG_ERROR_CC ( m_connection_context , " Maximum packet size exceed!, m_max_packet_size = " < < std : : min < size_t > ( max_packet_size , max_bytes )
< < " , packet header received " < < m_current_head . m_cb < < " , command " < < m_current_head . m_command
2014-03-03 22:07:58 +00:00
< < " , connection will be closed. " ) ;
return false ;
}
}
break ;
default :
LOG_ERROR_CC ( m_connection_context , " Undefined state in levin_server_impl::connection_handler, m_state= " < < m_state ) ;
return false ;
}
}
return true ;
}
bool after_init_connection ( )
{
if ( ! m_connection_initialized )
{
m_connection_initialized = true ;
m_config . add_connection ( this ) ;
}
return true ;
}
template < class callback_t >
2021-01-19 02:22:32 +00:00
bool async_invoke ( int command , message_writer in_msg , const callback_t & cb , size_t timeout = LEVIN_DEFAULT_TIMEOUT_PRECONFIGURED )
2014-03-03 22:07:58 +00:00
{
misc_utils : : auto_scope_leave_caller scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler : : finish_outer_call , this ) ) ;
if ( timeout = = LEVIN_DEFAULT_TIMEOUT_PRECONFIGURED )
timeout = m_config . m_invoke_timeout ;
int err_code = LEVIN_OK ;
do
{
CRITICAL_REGION_LOCAL ( m_call_lock ) ;
boost : : interprocess : : ipcdetail : : atomic_write32 ( & m_invoke_buf_ready , 0 ) ;
2019-05-11 15:38:35 +00:00
CRITICAL_REGION_BEGIN ( m_invoke_response_handlers_lock ) ;
2014-03-03 22:07:58 +00:00
2020-12-30 00:58:53 +00:00
if ( command = = m_connection_context . handshake_command ( ) )
m_max_packet_size = m_config . m_max_packet_size ;
2021-01-19 02:22:32 +00:00
if ( ! send_message ( in_msg . finalize_invoke ( command ) ) )
2014-03-03 22:07:58 +00:00
{
LOG_ERROR_CC ( m_connection_context , " Failed to do_send " ) ;
err_code = LEVIN_ERROR_CONNECTION ;
break ;
}
if ( ! add_invoke_response_handler ( cb , timeout , * this , command ) )
{
err_code = LEVIN_ERROR_CONNECTION_DESTROYED ;
break ;
}
CRITICAL_REGION_END ( ) ;
} while ( false ) ;
if ( LEVIN_OK ! = err_code )
{
2019-05-16 20:34:22 +00:00
epee : : span < const uint8_t > stub_buff = nullptr ;
2014-03-03 22:07:58 +00:00
// Never call callback inside critical section, that can cause deadlock
cb ( err_code , stub_buff , m_connection_context ) ;
return false ;
}
return true ;
}
2021-01-19 02:22:32 +00:00
int invoke ( int command , message_writer in_msg , std : : string & buff_out )
2014-03-03 22:07:58 +00:00
{
misc_utils : : auto_scope_leave_caller scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler : : finish_outer_call , this ) ) ;
CRITICAL_REGION_LOCAL ( m_call_lock ) ;
boost : : interprocess : : ipcdetail : : atomic_write32 ( & m_invoke_buf_ready , 0 ) ;
2020-12-30 00:58:53 +00:00
if ( command = = m_connection_context . handshake_command ( ) )
m_max_packet_size = m_config . m_max_packet_size ;
2021-01-19 02:22:32 +00:00
if ( ! send_message ( in_msg . finalize_invoke ( command ) ) )
2014-03-03 22:07:58 +00:00
{
2019-05-11 15:38:35 +00:00
LOG_ERROR_CC ( m_connection_context , " Failed to send request " ) ;
2014-03-03 22:07:58 +00:00
return LEVIN_ERROR_CONNECTION ;
}
uint64_t ticks_start = misc_utils : : get_tick_count ( ) ;
2017-06-04 21:36:09 +00:00
size_t prev_size = 0 ;
2014-03-03 22:07:58 +00:00
2021-01-21 20:14:36 +00:00
while ( ! boost : : interprocess : : ipcdetail : : atomic_read32 ( & m_invoke_buf_ready ) & & ! m_protocol_released )
2014-03-03 22:07:58 +00:00
{
2017-06-04 21:36:09 +00:00
if ( m_cache_in_buffer . size ( ) - prev_size > = MIN_BYTES_WANTED )
{
prev_size = m_cache_in_buffer . size ( ) ;
ticks_start = misc_utils : : get_tick_count ( ) ;
}
2014-03-03 22:07:58 +00:00
if ( misc_utils : : get_tick_count ( ) - ticks_start > m_config . m_invoke_timeout )
{
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 16:34:23 +00:00
MWARNING ( m_connection_context < < " invoke timeout ( " < < m_config . m_invoke_timeout < < " ), closing connection " ) ;
2014-03-03 22:07:58 +00:00
close ( ) ;
return LEVIN_ERROR_CONNECTION_TIMEDOUT ;
}
if ( ! m_pservice_endpoint - > call_run_once_service_io ( ) )
return LEVIN_ERROR_CONNECTION_DESTROYED ;
}
2021-01-21 20:14:36 +00:00
if ( m_protocol_released )
2014-03-03 22:07:58 +00:00
return LEVIN_ERROR_CONNECTION_DESTROYED ;
CRITICAL_REGION_BEGIN ( m_local_inv_buff_lock ) ;
buff_out . swap ( m_local_inv_buff ) ;
m_local_inv_buff . clear ( ) ;
CRITICAL_REGION_END ( ) ;
return m_invoke_result_code ;
}
2021-01-19 02:22:32 +00:00
/*! Sends `message` without adding a levin header. The message must have been
created with ` make_noise_notify ` , ` make_fragmented_notify ` , or
` message_writer : : finalize_notify ` . See additional instructions for
2019-05-16 20:34:22 +00:00
` make_fragmented_notify ` .
\ return 1 on success */
int send ( byte_slice message )
{
const misc_utils : : auto_scope_leave_caller scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler : : finish_outer_call , this )
) ;
2021-01-19 02:22:32 +00:00
if ( ! send_message ( std : : move ( message ) ) )
2019-05-16 20:34:22 +00:00
{
LOG_ERROR_CC ( m_connection_context , " Failed to send message, dropping it " ) ;
return - 1 ;
}
return 1 ;
}
2014-03-03 22:07:58 +00:00
//------------------------------------------------------------------------------------------
boost : : uuids : : uuid get_connection_id ( ) { return m_connection_context . m_connection_id ; }
//------------------------------------------------------------------------------------------
t_connection_context & get_context_ref ( ) { return m_connection_context ; }
} ;
//------------------------------------------------------------------------------------------
template < class t_connection_context >
void async_protocol_handler_config < t_connection_context > : : del_connection ( async_protocol_handler < t_connection_context > * pconn )
{
CRITICAL_REGION_BEGIN ( m_connects_lock ) ;
m_connects . erase ( pconn - > get_connection_id ( ) ) ;
CRITICAL_REGION_END ( ) ;
m_pcommands_handler - > on_connection_close ( pconn - > m_connection_context ) ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2018-01-20 21:44:23 +00:00
void async_protocol_handler_config < t_connection_context > : : delete_connections ( size_t count , bool incoming )
{
2021-03-12 11:59:06 +00:00
std : : vector < typename connections_map : : mapped_type > connections ;
auto scope_exit_handler = misc_utils : : create_scope_leave_handler ( [ & connections ] {
for ( auto & aph : connections )
aph - > finish_outer_call ( ) ;
} ) ;
2018-01-20 21:44:23 +00:00
CRITICAL_REGION_BEGIN ( m_connects_lock ) ;
for ( auto & c : m_connects )
{
if ( c . second - > m_connection_context . m_is_income = = incoming )
2021-03-12 11:59:06 +00:00
if ( c . second - > start_outer_call ( ) )
connections . push_back ( c . second ) ;
2018-01-20 21:44:23 +00:00
}
// close random connections from the provided set
// TODO or better just keep removing random elements (performance)
unsigned seed = std : : chrono : : system_clock : : now ( ) . time_since_epoch ( ) . count ( ) ;
shuffle ( connections . begin ( ) , connections . end ( ) , std : : default_random_engine ( seed ) ) ;
2021-03-12 11:59:06 +00:00
for ( size_t i = 0 ; i < connections . size ( ) & & i < count ; + + i )
m_connects . erase ( connections [ i ] - > get_connection_id ( ) ) ;
2018-01-20 21:44:23 +00:00
CRITICAL_REGION_END ( ) ;
2021-03-12 11:59:06 +00:00
for ( size_t i = 0 ; i < connections . size ( ) & & i < count ; + + i )
connections [ i ] - > close ( ) ;
2018-01-20 21:44:23 +00:00
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2015-02-12 19:59:39 +00:00
void async_protocol_handler_config < t_connection_context > : : del_out_connections ( size_t count )
2015-01-05 19:30:17 +00:00
{
2018-01-20 21:44:23 +00:00
delete_connections ( count , false ) ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
void async_protocol_handler_config < t_connection_context > : : del_in_connections ( size_t count )
{
delete_connections ( count , true ) ;
2015-01-05 19:30:17 +00:00
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2014-03-03 22:07:58 +00:00
void async_protocol_handler_config < t_connection_context > : : add_connection ( async_protocol_handler < t_connection_context > * pconn )
{
CRITICAL_REGION_BEGIN ( m_connects_lock ) ;
m_connects [ pconn - > get_connection_id ( ) ] = pconn ;
CRITICAL_REGION_END ( ) ;
m_pcommands_handler - > on_connection_new ( pconn - > m_connection_context ) ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
async_protocol_handler < t_connection_context > * async_protocol_handler_config < t_connection_context > : : find_connection ( boost : : uuids : : uuid connection_id ) const
{
auto it = m_connects . find ( connection_id ) ;
return it = = m_connects . end ( ) ? 0 : it - > second ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
int async_protocol_handler_config < t_connection_context > : : find_and_lock_connection ( boost : : uuids : : uuid connection_id , async_protocol_handler < t_connection_context > * & aph )
{
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
aph = find_connection ( connection_id ) ;
if ( 0 = = aph )
return LEVIN_ERROR_CONNECTION_NOT_FOUND ;
if ( ! aph - > start_outer_call ( ) )
return LEVIN_ERROR_CONNECTION_DESTROYED ;
return LEVIN_OK ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2021-01-19 02:22:32 +00:00
int async_protocol_handler_config < t_connection_context > : : invoke ( int command , message_writer in_msg , std : : string & buff_out , boost : : uuids : : uuid connection_id )
2014-03-03 22:07:58 +00:00
{
async_protocol_handler < t_connection_context > * aph ;
int r = find_and_lock_connection ( connection_id , aph ) ;
2021-01-19 02:22:32 +00:00
return LEVIN_OK = = r ? aph - > invoke ( command , std : : move ( in_msg ) , buff_out ) : r ;
2014-03-03 22:07:58 +00:00
}
//------------------------------------------------------------------------------------------
template < class t_connection_context > template < class callback_t >
2021-01-19 02:22:32 +00:00
int async_protocol_handler_config < t_connection_context > : : invoke_async ( int command , message_writer in_msg , boost : : uuids : : uuid connection_id , const callback_t & cb , size_t timeout )
2014-03-03 22:07:58 +00:00
{
async_protocol_handler < t_connection_context > * aph ;
int r = find_and_lock_connection ( connection_id , aph ) ;
2021-01-19 02:22:32 +00:00
return LEVIN_OK = = r ? aph - > async_invoke ( command , std : : move ( in_msg ) , cb , timeout ) : r ;
2014-03-03 22:07:58 +00:00
}
//------------------------------------------------------------------------------------------
template < class t_connection_context > template < class callback_t >
2018-02-02 18:45:12 +00:00
bool async_protocol_handler_config < t_connection_context > : : foreach_connection ( const callback_t & cb )
2014-03-03 22:07:58 +00:00
{
2021-01-19 17:09:32 +00:00
std : : vector < typename connections_map : : mapped_type > conn ;
auto scope_exit_handler = misc_utils : : create_scope_leave_handler ( [ & conn ] {
for ( auto & aph : conn )
aph - > finish_outer_call ( ) ;
} ) ;
2021-03-12 11:59:06 +00:00
CRITICAL_REGION_BEGIN ( m_connects_lock ) ;
conn . reserve ( m_connects . size ( ) ) ;
2021-01-19 17:09:32 +00:00
for ( auto & e : m_connects )
if ( e . second - > start_outer_call ( ) )
conn . push_back ( e . second ) ;
2021-03-12 11:59:06 +00:00
CRITICAL_REGION_END ( )
2021-01-19 17:09:32 +00:00
for ( auto & aph : conn )
if ( ! cb ( aph - > get_context_ref ( ) ) )
2014-03-03 22:07:58 +00:00
return false ;
2021-01-19 17:09:32 +00:00
2014-03-03 22:07:58 +00:00
return true ;
}
//------------------------------------------------------------------------------------------
2017-07-02 21:41:15 +00:00
template < class t_connection_context > template < class callback_t >
2018-02-02 18:45:12 +00:00
bool async_protocol_handler_config < t_connection_context > : : for_connection ( const boost : : uuids : : uuid & connection_id , const callback_t & cb )
2017-07-02 21:41:15 +00:00
{
2021-03-12 11:59:06 +00:00
async_protocol_handler < t_connection_context > * aph = nullptr ;
if ( find_and_lock_connection ( connection_id , aph ) ! = LEVIN_OK )
2021-01-19 17:09:32 +00:00
return false ;
auto scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler < t_connection_context > : : finish_outer_call , aph ) ) ;
2017-07-02 21:41:15 +00:00
if ( ! cb ( aph - > get_context_ref ( ) ) )
return false ;
return true ;
}
//------------------------------------------------------------------------------------------
2014-03-03 22:07:58 +00:00
template < class t_connection_context >
size_t async_protocol_handler_config < t_connection_context > : : get_connections_count ( )
{
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
return m_connects . size ( ) ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2019-06-18 22:47:05 +00:00
size_t async_protocol_handler_config < t_connection_context > : : get_out_connections_count ( )
{
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
size_t count = 0 ;
for ( const auto & c : m_connects )
if ( ! c . second - > m_connection_context . m_is_income )
+ + count ;
return count ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
size_t async_protocol_handler_config < t_connection_context > : : get_in_connections_count ( )
{
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
size_t count = 0 ;
for ( const auto & c : m_connects )
if ( c . second - > m_connection_context . m_is_income )
+ + count ;
return count ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2017-10-09 15:46:42 +00:00
void async_protocol_handler_config < t_connection_context > : : set_handler ( levin_commands_handler < t_connection_context > * handler , void ( * destroy ) ( levin_commands_handler < t_connection_context > * ) )
{
if ( m_pcommands_handler & & m_pcommands_handler_destroy )
( * m_pcommands_handler_destroy ) ( m_pcommands_handler ) ;
m_pcommands_handler = handler ;
m_pcommands_handler_destroy = destroy ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2019-05-16 20:34:22 +00:00
int async_protocol_handler_config < t_connection_context > : : send ( byte_slice message , const boost : : uuids : : uuid & connection_id )
{
async_protocol_handler < t_connection_context > * aph ;
int r = find_and_lock_connection ( connection_id , aph ) ;
return LEVIN_OK = = r ? aph - > send ( std : : move ( message ) ) : 0 ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
2014-03-03 22:07:58 +00:00
bool async_protocol_handler_config < t_connection_context > : : close ( boost : : uuids : : uuid connection_id )
{
2021-03-12 11:59:06 +00:00
async_protocol_handler < t_connection_context > * aph = nullptr ;
if ( find_and_lock_connection ( connection_id , aph ) ! = LEVIN_OK )
2019-12-30 17:24:42 +00:00
return false ;
2021-03-12 11:59:06 +00:00
auto scope_exit_handler = misc_utils : : create_scope_leave_handler (
boost : : bind ( & async_protocol_handler < t_connection_context > : : finish_outer_call , aph ) ) ;
2019-12-30 17:24:42 +00:00
if ( ! aph - > close ( ) )
return false ;
2021-03-12 11:59:06 +00:00
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
2019-12-30 17:24:42 +00:00
m_connects . erase ( connection_id ) ;
return true ;
2014-03-03 22:07:58 +00:00
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
bool async_protocol_handler_config < t_connection_context > : : update_connection_context ( const t_connection_context & contxt )
{
CRITICAL_REGION_LOCAL ( m_connects_lock ) ;
async_protocol_handler < t_connection_context > * aph = find_connection ( contxt . m_connection_id ) ;
if ( 0 = = aph )
return false ;
aph - > update_connection_context ( contxt ) ;
return true ;
}
//------------------------------------------------------------------------------------------
template < class t_connection_context >
bool async_protocol_handler_config < t_connection_context > : : request_callback ( boost : : uuids : : uuid connection_id )
{
async_protocol_handler < t_connection_context > * aph ;
int r = find_and_lock_connection ( connection_id , aph ) ;
if ( LEVIN_OK = = r )
{
aph - > request_callback ( ) ;
return true ;
}
else
{
return false ;
}
}
}
}