mirror of
https://github.com/transmission/transmission
synced 2024-12-27 01:57:52 +00:00
1771 lines
50 KiB
C
1771 lines
50 KiB
C
/*
|
|
* This file Copyright (C) 2007 Charles Kerr <charles@rebelbase.com>
|
|
*
|
|
* This file is licensed by the GPL version 2. Works owned by the
|
|
* Transmission project are granted a special exemption to clause 2(b)
|
|
* so that the bulk of its code can remain under the MIT license.
|
|
* This exemption does not extend to derived works not owned by
|
|
* the Transmission project.
|
|
*
|
|
* $Id$
|
|
*/
|
|
|
|
#include <assert.h>
|
|
#include <ctype.h>
|
|
#include <errno.h>
|
|
#include <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
#include <libgen.h> /* basename */
|
|
|
|
#include <netinet/in.h> /* struct in_addr */
|
|
|
|
#include <sys/types.h> /* event.h needs this */
|
|
#include <event.h>
|
|
|
|
#include "transmission.h"
|
|
#include "bencode.h"
|
|
#include "completion.h"
|
|
#include "inout.h"
|
|
#include "list.h"
|
|
#include "peer-io.h"
|
|
#include "peer-mgr.h"
|
|
#include "peer-mgr-private.h"
|
|
#include "peer-msgs.h"
|
|
#include "ratecontrol.h"
|
|
#include "trevent.h"
|
|
#include "utils.h"
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
#define MAX_ALLOWED_SET_COUNT 10 /* number of pieces generated for allow-fast,
|
|
threshold for fast-allowing others */
|
|
|
|
enum
|
|
{
|
|
BT_CHOKE = 0,
|
|
BT_UNCHOKE = 1,
|
|
BT_INTERESTED = 2,
|
|
BT_NOT_INTERESTED = 3,
|
|
BT_HAVE = 4,
|
|
BT_BITFIELD = 5,
|
|
BT_REQUEST = 6,
|
|
BT_PIECE = 7,
|
|
BT_CANCEL = 8,
|
|
BT_PORT = 9,
|
|
BT_SUGGEST = 13,
|
|
BT_HAVE_ALL = 14,
|
|
BT_HAVE_NONE = 15,
|
|
BT_REJECT = 16,
|
|
BT_ALLOWED_FAST = 17,
|
|
BT_LTEP = 20,
|
|
|
|
LTEP_HANDSHAKE = 0,
|
|
|
|
OUR_LTEP_PEX = 1,
|
|
|
|
MAX_REQUEST_BYTE_COUNT = (16 * 1024), /* drop requests who want too much */
|
|
|
|
KEEPALIVE_INTERVAL_SECS = 90, /* idle seconds before we send a keepalive */
|
|
PEX_INTERVAL = (60 * 1000), /* msec between calls to sendPex() */
|
|
PEER_PULSE_INTERVAL = (100), /* msec between calls to pulse() */
|
|
RATE_PULSE_INTERVAL = (333), /* msec between calls to ratePulse() */
|
|
};
|
|
|
|
enum
|
|
{
|
|
AWAITING_BT_LENGTH,
|
|
AWAITING_BT_MESSAGE,
|
|
READING_BT_PIECE
|
|
};
|
|
|
|
struct peer_request
|
|
{
|
|
uint32_t index;
|
|
uint32_t offset;
|
|
uint32_t length;
|
|
time_t time_requested;
|
|
};
|
|
|
|
static int
|
|
compareRequest( const void * va, const void * vb )
|
|
{
|
|
struct peer_request * a = (struct peer_request*) va;
|
|
struct peer_request * b = (struct peer_request*) vb;
|
|
if( a->index != b->index ) return a->index - b->index;
|
|
if( a->offset != b->offset ) return a->offset - b->offset;
|
|
if( a->length != b->length ) return a->length - b->length;
|
|
return 0;
|
|
}
|
|
|
|
struct tr_peermsgs
|
|
{
|
|
tr_peer * info;
|
|
|
|
tr_handle * handle;
|
|
tr_torrent * torrent;
|
|
tr_peerIo * io;
|
|
|
|
tr_publisher_t * publisher;
|
|
|
|
struct evbuffer * outBlock; /* buffer of all the current piece message */
|
|
struct evbuffer * outMessages; /* buffer of all the non-piece messages */
|
|
struct evbuffer * inBlock; /* the block we're currently receiving */
|
|
tr_list * peerAskedFor;
|
|
tr_list * peerAskedForFast;
|
|
tr_list * clientAskedFor;
|
|
tr_list * clientWillAskFor;
|
|
|
|
tr_timer * rateTimer;
|
|
tr_timer * pulseTimer;
|
|
tr_timer * pexTimer;
|
|
|
|
struct peer_request blockToUs; /* the block currntly being sent to us */
|
|
|
|
time_t lastReqAddedAt;
|
|
time_t clientSentPexAt;
|
|
time_t clientSentAnythingAt;
|
|
|
|
unsigned int notListening : 1;
|
|
unsigned int peerSupportsPex : 1;
|
|
unsigned int clientSentLtepHandshake : 1;
|
|
unsigned int peerSentLtepHandshake : 1;
|
|
|
|
tr_bitfield * clientAllowedPieces;
|
|
tr_bitfield * peerAllowedPieces;
|
|
|
|
uint8_t state;
|
|
uint8_t ut_pex_id;
|
|
uint16_t pexCount;
|
|
uint32_t incomingMessageLength;
|
|
uint32_t maxActiveRequests;
|
|
uint32_t minActiveRequests;
|
|
|
|
tr_pex * pex;
|
|
};
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
static void
|
|
myDebug( const char * file, int line,
|
|
const struct tr_peermsgs * msgs,
|
|
const char * fmt, ... )
|
|
{
|
|
FILE * fp = tr_getLog( );
|
|
if( fp != NULL )
|
|
{
|
|
va_list args;
|
|
char timestr[64];
|
|
struct evbuffer * buf = evbuffer_new( );
|
|
char * myfile = tr_strdup( file );
|
|
|
|
evbuffer_add_printf( buf, "[%s] %s [%s]: ",
|
|
tr_getLogTimeStr( timestr, sizeof(timestr) ),
|
|
tr_peerIoGetAddrStr( msgs->io ),
|
|
msgs->info->client );
|
|
va_start( args, fmt );
|
|
evbuffer_add_vprintf( buf, fmt, args );
|
|
va_end( args );
|
|
evbuffer_add_printf( buf, " (%s:%d)\n", basename(myfile), line );
|
|
fwrite( EVBUFFER_DATA(buf), 1, EVBUFFER_LENGTH(buf), fp );
|
|
|
|
tr_free( myfile );
|
|
evbuffer_free( buf );
|
|
}
|
|
}
|
|
|
|
#define dbgmsg(msgs, fmt...) myDebug(__FILE__, __LINE__, msgs, ##fmt )
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
static void
|
|
protocolSendRequest( tr_peermsgs * msgs, const struct peer_request * req )
|
|
{
|
|
tr_peerIo * io = msgs->io;
|
|
struct evbuffer * out = msgs->outMessages;
|
|
|
|
dbgmsg( msgs, "requesting %u:%u->%u", req->index, req->offset, req->length );
|
|
tr_peerIoWriteUint32( io, out, sizeof(uint8_t) + 3*sizeof(uint32_t) );
|
|
tr_peerIoWriteUint8 ( io, out, BT_REQUEST );
|
|
tr_peerIoWriteUint32( io, out, req->index );
|
|
tr_peerIoWriteUint32( io, out, req->offset );
|
|
tr_peerIoWriteUint32( io, out, req->length );
|
|
}
|
|
|
|
static void
|
|
protocolSendCancel( tr_peermsgs * msgs, const struct peer_request * req )
|
|
{
|
|
tr_peerIo * io = msgs->io;
|
|
struct evbuffer * out = msgs->outMessages;
|
|
|
|
dbgmsg( msgs, "cancelling %u:%u->%u", req->index, req->offset, req->length );
|
|
tr_peerIoWriteUint32( io, out, sizeof(uint8_t) + 3*sizeof(uint32_t) );
|
|
tr_peerIoWriteUint8 ( io, out, BT_CANCEL );
|
|
tr_peerIoWriteUint32( io, out, req->index );
|
|
tr_peerIoWriteUint32( io, out, req->offset );
|
|
tr_peerIoWriteUint32( io, out, req->length );
|
|
}
|
|
|
|
static void
|
|
protocolSendHave( tr_peermsgs * msgs, uint32_t index )
|
|
{
|
|
tr_peerIo * io = msgs->io;
|
|
struct evbuffer * out = msgs->outMessages;
|
|
|
|
dbgmsg( msgs, "sending Have %u", index );
|
|
tr_peerIoWriteUint32( io, out, sizeof(uint8_t) + sizeof(uint32_t) );
|
|
tr_peerIoWriteUint8 ( io, out, BT_HAVE );
|
|
tr_peerIoWriteUint32( io, out, index );
|
|
}
|
|
|
|
static void
|
|
protocolSendChoke( tr_peermsgs * msgs, int choke )
|
|
{
|
|
tr_peerIo * io = msgs->io;
|
|
struct evbuffer * out = msgs->outMessages;
|
|
|
|
dbgmsg( msgs, "sending %s", (choke ? "Choke" : "Unchoke") );
|
|
tr_peerIoWriteUint32( io, out, sizeof(uint8_t) );
|
|
tr_peerIoWriteUint8 ( io, out, choke ? BT_CHOKE : BT_UNCHOKE );
|
|
}
|
|
|
|
/**
|
|
*** EVENTS
|
|
**/
|
|
|
|
static const tr_peermsgs_event blankEvent = { 0, 0, 0, 0, 0.0f };
|
|
|
|
static void
|
|
publish( tr_peermsgs * msgs, tr_peermsgs_event * e )
|
|
{
|
|
tr_publisherPublish( msgs->publisher, msgs->info, e );
|
|
}
|
|
|
|
static void
|
|
fireGotError( tr_peermsgs * msgs )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_GOT_ERROR;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
static void
|
|
fireNeedReq( tr_peermsgs * msgs )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_NEED_REQ;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
static void
|
|
firePeerProgress( tr_peermsgs * msgs )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_PEER_PROGRESS;
|
|
e.progress = msgs->info->progress;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
static void
|
|
fireClientHave( tr_peermsgs * msgs, uint32_t pieceIndex )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_CLIENT_HAVE;
|
|
e.pieceIndex = pieceIndex;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
static void
|
|
fireGotBlock( tr_peermsgs * msgs, uint32_t pieceIndex, uint32_t offset, uint32_t length )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_CLIENT_BLOCK;
|
|
e.pieceIndex = pieceIndex;
|
|
e.offset = offset;
|
|
e.length = length;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
static void
|
|
fireCancelledReq( tr_peermsgs * msgs, const struct peer_request * req )
|
|
{
|
|
tr_peermsgs_event e = blankEvent;
|
|
e.eventType = TR_PEERMSG_CANCEL;
|
|
e.pieceIndex = req->index;
|
|
e.offset = req->offset;
|
|
e.length = req->length;
|
|
publish( msgs, &e );
|
|
}
|
|
|
|
/**
|
|
*** INTEREST
|
|
**/
|
|
|
|
static int
|
|
isPieceInteresting( const tr_peermsgs * peer,
|
|
int piece )
|
|
{
|
|
const tr_torrent * torrent = peer->torrent;
|
|
if( torrent->info.pieces[piece].dnd ) /* we don't want it */
|
|
return FALSE;
|
|
if( tr_cpPieceIsComplete( torrent->completion, piece ) ) /* we have it */
|
|
return FALSE;
|
|
if( !tr_bitfieldHas( peer->info->have, piece ) ) /* peer doesn't have it */
|
|
return FALSE;
|
|
if( tr_bitfieldHas( peer->info->banned, piece ) ) /* peer is banned */
|
|
return FALSE;
|
|
return TRUE;
|
|
}
|
|
|
|
/* "interested" means we'll ask for piece data if they unchoke us */
|
|
static int
|
|
isPeerInteresting( const tr_peermsgs * msgs )
|
|
{
|
|
int i;
|
|
const tr_torrent * torrent;
|
|
const tr_bitfield * bitfield;
|
|
const int clientIsSeed = tr_torrentIsSeed( msgs->torrent );
|
|
|
|
if( clientIsSeed )
|
|
return FALSE;
|
|
|
|
torrent = msgs->torrent;
|
|
bitfield = tr_cpPieceBitfield( torrent->completion );
|
|
|
|
if( !msgs->info->have )
|
|
return TRUE;
|
|
|
|
assert( bitfield->len == msgs->info->have->len );
|
|
for( i=0; i<torrent->info.pieceCount; ++i )
|
|
if( isPieceInteresting( msgs, i ) )
|
|
return TRUE;
|
|
|
|
return FALSE;
|
|
}
|
|
|
|
static void
|
|
sendInterest( tr_peermsgs * msgs, int weAreInterested )
|
|
{
|
|
assert( msgs != NULL );
|
|
assert( weAreInterested==0 || weAreInterested==1 );
|
|
|
|
msgs->info->clientIsInterested = weAreInterested;
|
|
dbgmsg( msgs, "Sending %s",
|
|
weAreInterested ? "Interested" : "Not Interested");
|
|
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, sizeof(uint8_t) );
|
|
tr_peerIoWriteUint8 ( msgs->io, msgs->outMessages,
|
|
weAreInterested ? BT_INTERESTED : BT_NOT_INTERESTED );
|
|
}
|
|
|
|
static void
|
|
updateInterest( tr_peermsgs * msgs )
|
|
{
|
|
const int i = isPeerInteresting( msgs );
|
|
if( i != msgs->info->clientIsInterested )
|
|
sendInterest( msgs, i );
|
|
if( i )
|
|
fireNeedReq( msgs );
|
|
}
|
|
|
|
#define MIN_CHOKE_PERIOD_SEC 10
|
|
|
|
static void
|
|
cancelAllRequestsToClientExceptFast( tr_peermsgs * msgs )
|
|
{
|
|
tr_list_free( &msgs->peerAskedFor, tr_free );
|
|
}
|
|
|
|
void
|
|
tr_peerMsgsSetChoke( tr_peermsgs * msgs, int choke )
|
|
{
|
|
const time_t fibrillationTime = time(NULL) - MIN_CHOKE_PERIOD_SEC;
|
|
|
|
assert( msgs != NULL );
|
|
assert( msgs->info != NULL );
|
|
assert( choke==0 || choke==1 );
|
|
|
|
if( msgs->info->chokeChangedAt > fibrillationTime )
|
|
{
|
|
dbgmsg( msgs, "Not changing choke to %d to avoid fibrillation", choke );
|
|
}
|
|
else if( msgs->info->peerIsChoked != choke )
|
|
{
|
|
msgs->info->peerIsChoked = choke;
|
|
if( choke )
|
|
cancelAllRequestsToClientExceptFast( msgs );
|
|
protocolSendChoke( msgs, choke );
|
|
msgs->info->chokeChangedAt = time( NULL );
|
|
}
|
|
}
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
void
|
|
tr_peerMsgsHave( tr_peermsgs * msgs,
|
|
uint32_t index )
|
|
{
|
|
protocolSendHave( msgs, index );
|
|
|
|
/* since we have more pieces now, we might not be interested in this peer */
|
|
updateInterest( msgs );
|
|
}
|
|
#if 0
|
|
static void
|
|
sendFastSuggest( tr_peermsgs * msgs,
|
|
uint32_t pieceIndex )
|
|
{
|
|
dbgmsg( msgs, "w00t SUGGESTing them piece #%d", pieceIndex );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, sizeof(uint8_t) + sizeof(uint32_t) );
|
|
tr_peerIoWriteUint8( msgs->io, msgs->outMessages, BT_SUGGEST );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, pieceIndex );
|
|
|
|
updateInterest( msgs );
|
|
}
|
|
#endif
|
|
static void
|
|
sendFastHave( tr_peermsgs * msgs,
|
|
int all)
|
|
{
|
|
dbgmsg( msgs, "w00t telling them we %s pieces", (all ? "HAVE_ALL" : "HAVE_NONE" ) );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, sizeof(uint8_t) );
|
|
tr_peerIoWriteUint8( msgs->io, msgs->outMessages, ( all ? BT_HAVE_ALL : BT_HAVE_NONE ) );
|
|
|
|
updateInterest( msgs );
|
|
}
|
|
|
|
static void
|
|
sendFastReject( tr_peermsgs * msgs,
|
|
uint32_t pieceIndex,
|
|
uint32_t offset,
|
|
uint32_t length )
|
|
{
|
|
assert( msgs != NULL );
|
|
assert( length > 0 );
|
|
|
|
if( tr_peerIoSupportsFEXT( msgs->io ) )
|
|
{
|
|
const uint32_t len = sizeof(uint8_t) + 3 * sizeof(uint32_t);
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, len );
|
|
tr_peerIoWriteUint8( msgs->io, msgs->outMessages, BT_REJECT );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, pieceIndex );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, offset );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, length );
|
|
}
|
|
}
|
|
|
|
static void
|
|
sendFastAllowed( tr_peermsgs * msgs,
|
|
uint32_t pieceIndex)
|
|
{
|
|
dbgmsg( msgs, "w00t telling them we ALLOW_FAST piece #%d", pieceIndex );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, sizeof(uint8_t) + sizeof(uint32_t) );
|
|
tr_peerIoWriteUint8( msgs->io, msgs->outMessages, BT_ALLOWED_FAST );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, pieceIndex );
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
sendFastAllowedSet( tr_peermsgs * msgs )
|
|
{
|
|
int i = 0;
|
|
while (i <= msgs->torrent->info.pieceCount )
|
|
{
|
|
if ( tr_bitfieldHas( msgs->peerAllowedPieces, i) )
|
|
sendFastAllowed( msgs, i );
|
|
i++;
|
|
}
|
|
}
|
|
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
static int
|
|
reqIsValid( const tr_peermsgs * msgs, uint32_t index, uint32_t offset, uint32_t length )
|
|
{
|
|
const tr_torrent * tor = msgs->torrent;
|
|
|
|
if( index >= (uint32_t) tor->info.pieceCount )
|
|
return FALSE;
|
|
if ( (int)offset >= tr_torPieceCountBytes( tor, (int)index ) )
|
|
return FALSE;
|
|
if( length > MAX_REQUEST_BYTE_COUNT )
|
|
return FALSE;
|
|
if( tr_pieceOffset( tor, index, offset, length ) > tor->info.totalSize )
|
|
return FALSE;
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static int
|
|
requestIsValid( const tr_peermsgs * msgs, const struct peer_request * req )
|
|
{
|
|
return reqIsValid( msgs, req->index, req->offset, req->length );
|
|
}
|
|
|
|
static void
|
|
pumpRequestQueue( tr_peermsgs * msgs )
|
|
{
|
|
const int max = msgs->maxActiveRequests;
|
|
const int min = msgs->minActiveRequests;
|
|
int count = tr_list_size( msgs->clientAskedFor );
|
|
int sent = 0;
|
|
|
|
if( count > min )
|
|
return;
|
|
if( msgs->info->clientIsChoked )
|
|
return;
|
|
|
|
while( ( count < max ) && ( msgs->clientWillAskFor != NULL ) )
|
|
{
|
|
struct peer_request * req = tr_list_pop_front( &msgs->clientWillAskFor );
|
|
protocolSendRequest( msgs, req );
|
|
req->time_requested = msgs->lastReqAddedAt = time( NULL );
|
|
tr_list_append( &msgs->clientAskedFor, req );
|
|
++count;
|
|
++sent;
|
|
}
|
|
|
|
if( sent )
|
|
dbgmsg( msgs, "pump sent %d requests, now have %d active and %d queued",
|
|
sent,
|
|
tr_list_size(msgs->clientAskedFor),
|
|
tr_list_size(msgs->clientWillAskFor) );
|
|
|
|
if( count < max )
|
|
fireNeedReq( msgs );
|
|
}
|
|
|
|
int
|
|
tr_peerMsgsAddRequest( tr_peermsgs * msgs,
|
|
uint32_t index,
|
|
uint32_t offset,
|
|
uint32_t length )
|
|
{
|
|
const int req_max = msgs->maxActiveRequests;
|
|
struct peer_request tmp, *req;
|
|
|
|
assert( msgs != NULL );
|
|
assert( msgs->torrent != NULL );
|
|
assert( reqIsValid( msgs, index, offset, length ) );
|
|
|
|
/**
|
|
*** Reasons to decline the request
|
|
**/
|
|
|
|
/* don't send requests to choked clients */
|
|
if( msgs->info->clientIsChoked ) {
|
|
dbgmsg( msgs, "declining request because they're choking us" );
|
|
return TR_ADDREQ_CLIENT_CHOKED;
|
|
}
|
|
|
|
/* peer doesn't have this piece */
|
|
if( !tr_bitfieldHas( msgs->info->have, index ) )
|
|
return TR_ADDREQ_MISSING;
|
|
|
|
/* peer's queue is full */
|
|
if( tr_list_size( msgs->clientWillAskFor ) >= req_max ) {
|
|
dbgmsg( msgs, "declining request because we're full" );
|
|
return TR_ADDREQ_FULL;
|
|
}
|
|
|
|
/* have we already asked for this piece? */
|
|
tmp.index = index;
|
|
tmp.offset = offset;
|
|
tmp.length = length;
|
|
if( tr_list_find( msgs->clientAskedFor, &tmp, compareRequest ) ) {
|
|
dbgmsg( msgs, "declining because it's a duplicate" );
|
|
return TR_ADDREQ_DUPLICATE;
|
|
}
|
|
if( tr_list_find( msgs->clientWillAskFor, &tmp, compareRequest ) ) {
|
|
dbgmsg( msgs, "declining because it's a duplicate" );
|
|
return TR_ADDREQ_DUPLICATE;
|
|
}
|
|
|
|
/**
|
|
*** Accept this request
|
|
**/
|
|
|
|
dbgmsg( msgs, "added req for piece %d, offset %d", (int)index, (int)offset );
|
|
req = tr_new0( struct peer_request, 1 );
|
|
*req = tmp;
|
|
tr_list_append( &msgs->clientWillAskFor, req );
|
|
return TR_ADDREQ_OK;
|
|
}
|
|
|
|
static void
|
|
cancelAllRequestsToPeer( tr_peermsgs * msgs )
|
|
{
|
|
struct peer_request * req;
|
|
|
|
while(( req = tr_list_pop_front( &msgs->clientWillAskFor ) ))
|
|
{
|
|
fireCancelledReq( msgs, req );
|
|
tr_free( req );
|
|
}
|
|
|
|
while(( req = tr_list_pop_front( &msgs->clientAskedFor ) ))
|
|
{
|
|
fireCancelledReq( msgs, req );
|
|
protocolSendCancel( msgs, req );
|
|
tr_free( req );
|
|
}
|
|
}
|
|
|
|
void
|
|
tr_peerMsgsCancel( tr_peermsgs * msgs,
|
|
uint32_t pieceIndex,
|
|
uint32_t offset,
|
|
uint32_t length )
|
|
{
|
|
struct peer_request *req, tmp;
|
|
|
|
assert( msgs != NULL );
|
|
assert( length > 0 );
|
|
|
|
/* have we asked the peer for this piece? */
|
|
tmp.index = pieceIndex;
|
|
tmp.offset = offset;
|
|
tmp.length = length;
|
|
|
|
/* if it's only in the queue and hasn't been sent yet, free it */
|
|
if(( req = tr_list_remove( &msgs->clientWillAskFor, &tmp, compareRequest ) ))
|
|
{
|
|
fireCancelledReq( msgs, req );
|
|
tr_free( req );
|
|
}
|
|
|
|
/* if it's already been sent, send a cancel message too */
|
|
if(( req = tr_list_remove( &msgs->clientAskedFor, &tmp, compareRequest ) ))
|
|
{
|
|
protocolSendCancel( msgs, req );
|
|
fireCancelledReq( msgs, req );
|
|
tr_free( req );
|
|
}
|
|
}
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
static void
|
|
sendLtepHandshake( tr_peermsgs * msgs )
|
|
{
|
|
benc_val_t val, *m;
|
|
char * buf;
|
|
int len;
|
|
int pex;
|
|
const char * v = TR_NAME " " USERAGENT_PREFIX;
|
|
const int port = tr_getPublicPort( msgs->handle );
|
|
struct evbuffer * outbuf;
|
|
|
|
if( msgs->clientSentLtepHandshake )
|
|
return;
|
|
|
|
outbuf = evbuffer_new( );
|
|
dbgmsg( msgs, "sending an ltep handshake" );
|
|
msgs->clientSentLtepHandshake = 1;
|
|
|
|
/* decide if we want to advertise pex support */
|
|
if( !tr_torrentIsPexEnabled( msgs->torrent ) )
|
|
pex = 0;
|
|
else if( msgs->peerSentLtepHandshake )
|
|
pex = msgs->peerSupportsPex ? 1 : 0;
|
|
else
|
|
pex = 1;
|
|
|
|
tr_bencInit( &val, TYPE_DICT );
|
|
tr_bencDictReserve( &val, 4 );
|
|
tr_bencInitInt( tr_bencDictAdd( &val, "e" ), 1 );
|
|
m = tr_bencDictAdd( &val, "m" );
|
|
tr_bencInit( m, TYPE_DICT );
|
|
if( pex ) {
|
|
tr_bencDictReserve( m, 1 );
|
|
tr_bencInitInt( tr_bencDictAdd( m, "ut_pex" ), OUR_LTEP_PEX );
|
|
}
|
|
if( port > 0 )
|
|
tr_bencInitInt( tr_bencDictAdd( &val, "p" ), port );
|
|
tr_bencInitStr( tr_bencDictAdd( &val, "v" ), v, 0, 1 );
|
|
buf = tr_bencSaveMalloc( &val, &len );
|
|
|
|
tr_peerIoWriteUint32( msgs->io, outbuf, 2*sizeof(uint8_t) + len );
|
|
tr_peerIoWriteUint8 ( msgs->io, outbuf, BT_LTEP );
|
|
tr_peerIoWriteUint8 ( msgs->io, outbuf, LTEP_HANDSHAKE );
|
|
tr_peerIoWriteBytes ( msgs->io, outbuf, buf, len );
|
|
|
|
tr_peerIoWriteBuf( msgs->io, outbuf );
|
|
|
|
#if 0
|
|
dbgmsg( msgs, "here is the ltep handshake we sent:" );
|
|
tr_bencPrint( &val );
|
|
#endif
|
|
|
|
/* cleanup */
|
|
tr_bencFree( &val );
|
|
tr_free( buf );
|
|
evbuffer_free( outbuf );
|
|
}
|
|
|
|
static void
|
|
parseLtepHandshake( tr_peermsgs * msgs, int len, struct evbuffer * inbuf )
|
|
{
|
|
benc_val_t val, * sub;
|
|
uint8_t * tmp = tr_new( uint8_t, len );
|
|
|
|
tr_peerIoReadBytes( msgs->io, inbuf, tmp, len );
|
|
msgs->peerSentLtepHandshake = 1;
|
|
|
|
if( tr_bencLoad( tmp, len, &val, NULL ) || val.type!=TYPE_DICT ) {
|
|
dbgmsg( msgs, "GET extended-handshake, couldn't get dictionary" );
|
|
tr_free( tmp );
|
|
return;
|
|
}
|
|
|
|
#if 0
|
|
dbgmsg( msgs, "here is the ltep handshake we read:" );
|
|
tr_bencPrint( &val );
|
|
#endif
|
|
|
|
/* does the peer prefer encrypted connections? */
|
|
sub = tr_bencDictFind( &val, "e" );
|
|
if( tr_bencIsInt( sub ) )
|
|
msgs->info->encryption_preference = sub->val.i
|
|
? ENCRYPTION_PREFERENCE_YES
|
|
: ENCRYPTION_PREFERENCE_NO;
|
|
|
|
/* check supported messages for utorrent pex */
|
|
sub = tr_bencDictFind( &val, "m" );
|
|
if( tr_bencIsDict( sub ) ) {
|
|
sub = tr_bencDictFind( sub, "ut_pex" );
|
|
if( tr_bencIsInt( sub ) ) {
|
|
msgs->peerSupportsPex = 1;
|
|
msgs->ut_pex_id = (uint8_t) sub->val.i;
|
|
dbgmsg( msgs, "msgs->ut_pex is %d", (int)msgs->ut_pex_id );
|
|
}
|
|
}
|
|
|
|
/* get peer's listening port */
|
|
sub = tr_bencDictFind( &val, "p" );
|
|
if( tr_bencIsInt( sub ) ) {
|
|
msgs->info->port = htons( (uint16_t)sub->val.i );
|
|
dbgmsg( msgs, "msgs->port is now %hu", msgs->info->port );
|
|
}
|
|
|
|
tr_bencFree( &val );
|
|
tr_free( tmp );
|
|
}
|
|
|
|
static void
|
|
parseUtPex( tr_peermsgs * msgs, int msglen, struct evbuffer * inbuf )
|
|
{
|
|
benc_val_t val, * sub;
|
|
uint8_t * tmp;
|
|
|
|
if( !tr_torrentIsPexEnabled( msgs->torrent ) ) /* no sharing! */
|
|
return;
|
|
|
|
tmp = tr_new( uint8_t, msglen );
|
|
tr_peerIoReadBytes( msgs->io, inbuf, tmp, msglen );
|
|
|
|
if( tr_bencLoad( tmp, msglen, &val, NULL ) || !tr_bencIsDict( &val ) ) {
|
|
dbgmsg( msgs, "GET can't read extended-pex dictionary" );
|
|
tr_free( tmp );
|
|
return;
|
|
}
|
|
|
|
sub = tr_bencDictFind( &val, "added" );
|
|
if( tr_bencIsStr(sub) && ((sub->val.s.i % 6) == 0)) {
|
|
const int n = sub->val.s.i / 6 ;
|
|
dbgmsg( msgs, "got %d peers from uT pex", n );
|
|
tr_peerMgrAddPeers( msgs->handle->peerMgr,
|
|
msgs->torrent->info.hash,
|
|
TR_PEER_FROM_PEX,
|
|
(uint8_t*)sub->val.s.s, n );
|
|
}
|
|
|
|
tr_bencFree( &val );
|
|
tr_free( tmp );
|
|
}
|
|
|
|
static void
|
|
sendPex( tr_peermsgs * msgs );
|
|
|
|
static void
|
|
parseLtep( tr_peermsgs * msgs, int msglen, struct evbuffer * inbuf )
|
|
{
|
|
uint8_t ltep_msgid;
|
|
|
|
tr_peerIoReadUint8( msgs->io, inbuf, <ep_msgid );
|
|
msglen--;
|
|
|
|
if( ltep_msgid == LTEP_HANDSHAKE )
|
|
{
|
|
dbgmsg( msgs, "got ltep handshake" );
|
|
parseLtepHandshake( msgs, msglen, inbuf );
|
|
sendLtepHandshake( msgs );
|
|
sendPex( msgs );
|
|
}
|
|
else if( ltep_msgid == msgs->ut_pex_id )
|
|
{
|
|
dbgmsg( msgs, "got ut pex" );
|
|
msgs->peerSupportsPex = 1;
|
|
parseUtPex( msgs, msglen, inbuf );
|
|
}
|
|
else
|
|
{
|
|
dbgmsg( msgs, "skipping unknown ltep message (%d)", (int)ltep_msgid );
|
|
evbuffer_drain( inbuf, msglen );
|
|
}
|
|
}
|
|
|
|
static int
|
|
readBtLength( tr_peermsgs * msgs, struct evbuffer * inbuf )
|
|
{
|
|
uint32_t len;
|
|
const size_t needlen = sizeof(uint32_t);
|
|
|
|
if( EVBUFFER_LENGTH(inbuf) < needlen )
|
|
return READ_MORE;
|
|
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &len );
|
|
|
|
if( len == 0 ) /* peer sent us a keepalive message */
|
|
dbgmsg( msgs, "got KeepAlive" );
|
|
else {
|
|
msgs->incomingMessageLength = len;
|
|
msgs->state = AWAITING_BT_MESSAGE;
|
|
}
|
|
|
|
return READ_AGAIN;
|
|
}
|
|
|
|
static void
|
|
updatePeerProgress( tr_peermsgs * msgs )
|
|
{
|
|
msgs->info->progress = tr_bitfieldCountTrueBits( msgs->info->have ) / (float)msgs->torrent->info.pieceCount;
|
|
dbgmsg( msgs, "peer progress is %f", msgs->info->progress );
|
|
updateInterest( msgs );
|
|
firePeerProgress( msgs );
|
|
}
|
|
|
|
static int
|
|
clientCanSendFastBlock( const tr_peermsgs * msgs UNUSED )
|
|
{
|
|
/* FIXME(tiennou): base this on how many blocks we've already sent this
|
|
* peer, or maybe how many fast blocks per minute we've sent overall,
|
|
* or maybe how much bandwidth we're already using up w/o fast peers.
|
|
* I don't know what the Right Thing here is, but
|
|
* the previous measurement of how many pieces we have is not enough. */
|
|
return FALSE;
|
|
}
|
|
|
|
static void
|
|
peerMadeRequest( tr_peermsgs * msgs, const struct peer_request * req )
|
|
{
|
|
const int reqIsValid = requestIsValid( msgs, req );
|
|
const int clientHasPiece = reqIsValid && tr_cpPieceIsComplete( msgs->torrent->completion, req->index );
|
|
const int peerIsChoked = msgs->info->peerIsChoked;
|
|
const int peerIsFast = tr_peerIoSupportsFEXT( msgs->io );
|
|
const int pieceIsFast = reqIsValid && tr_bitfieldHas( msgs->peerAllowedPieces, req->index );
|
|
const int canSendFast = clientCanSendFastBlock( msgs );
|
|
|
|
if( !reqIsValid ) /* bad request */
|
|
{
|
|
dbgmsg( msgs, "rejecting an invalid request." );
|
|
sendFastReject( msgs, req->index, req->offset, req->length );
|
|
}
|
|
else if( !clientHasPiece ) /* we don't have it */
|
|
{
|
|
dbgmsg( msgs, "rejecting request for a piece we don't have." );
|
|
sendFastReject( msgs, req->index, req->offset, req->length );
|
|
}
|
|
else if( peerIsChoked && !peerIsFast ) /* maybe he doesn't know he's choked? */
|
|
{
|
|
tr_peerMsgsSetChoke( msgs, 1 );
|
|
sendFastReject( msgs, req->index, req->offset, req->length );
|
|
}
|
|
else if( peerIsChoked && peerIsFast && ( !pieceIsFast || !canSendFast ) )
|
|
{
|
|
sendFastReject( msgs, req->index, req->offset, req->length );
|
|
}
|
|
else /* YAY */
|
|
{
|
|
struct peer_request * tmp = tr_new( struct peer_request, 1 );
|
|
*tmp = *req;
|
|
if( peerIsFast && pieceIsFast )
|
|
tr_list_append( &msgs->peerAskedForFast, tmp );
|
|
else
|
|
tr_list_append( &msgs->peerAskedFor, tmp );
|
|
}
|
|
}
|
|
|
|
static int
|
|
messageLengthIsCorrect( const tr_peermsgs * msg, uint8_t id, uint32_t len )
|
|
{
|
|
switch( id )
|
|
{
|
|
case BT_CHOKE:
|
|
case BT_UNCHOKE:
|
|
case BT_INTERESTED:
|
|
case BT_NOT_INTERESTED:
|
|
case BT_HAVE_ALL:
|
|
case BT_HAVE_NONE:
|
|
return len==1;
|
|
|
|
case BT_HAVE:
|
|
case BT_SUGGEST:
|
|
case BT_ALLOWED_FAST:
|
|
return len==5;
|
|
|
|
case BT_BITFIELD:
|
|
return len == (msg->torrent->info.pieceCount+7u)/8u + 1u;
|
|
|
|
case BT_REQUEST:
|
|
case BT_CANCEL:
|
|
case BT_REJECT:
|
|
return len==13;
|
|
|
|
case BT_PIECE:
|
|
return len>9 && len<=16393;
|
|
|
|
case BT_PORT:
|
|
return len==3;
|
|
|
|
case BT_LTEP:
|
|
return len >= 2;
|
|
|
|
default:
|
|
return FALSE;
|
|
}
|
|
}
|
|
|
|
|
|
static int
|
|
readBtMessage( tr_peermsgs * msgs, struct evbuffer * inbuf )
|
|
{
|
|
uint8_t id;
|
|
uint32_t ui32;
|
|
uint32_t msglen = msgs->incomingMessageLength;
|
|
|
|
if( EVBUFFER_LENGTH(inbuf) < msglen )
|
|
return READ_MORE;
|
|
|
|
tr_peerIoReadUint8( msgs->io, inbuf, &id );
|
|
dbgmsg( msgs, "got BT id %d, len %d", (int)id, (int)msglen );
|
|
|
|
if( !messageLengthIsCorrect( msgs, id, msglen ) )
|
|
{
|
|
fireGotError( msgs );
|
|
return READ_DONE;
|
|
}
|
|
|
|
--msglen;
|
|
|
|
switch( id )
|
|
{
|
|
case BT_CHOKE:
|
|
dbgmsg( msgs, "got Choke" );
|
|
msgs->info->clientIsChoked = 1;
|
|
cancelAllRequestsToPeer( msgs );
|
|
cancelAllRequestsToClientExceptFast( msgs );
|
|
break;
|
|
|
|
case BT_UNCHOKE:
|
|
dbgmsg( msgs, "got Unchoke" );
|
|
msgs->info->clientIsChoked = 0;
|
|
fireNeedReq( msgs );
|
|
break;
|
|
|
|
case BT_INTERESTED:
|
|
dbgmsg( msgs, "got Interested" );
|
|
msgs->info->peerIsInterested = 1;
|
|
tr_peerMsgsSetChoke( msgs, 0 );
|
|
break;
|
|
|
|
case BT_NOT_INTERESTED:
|
|
dbgmsg( msgs, "got Not Interested" );
|
|
msgs->info->peerIsInterested = 0;
|
|
break;
|
|
|
|
case BT_HAVE:
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &ui32 );
|
|
dbgmsg( msgs, "got Have: %u", ui32 );
|
|
tr_bitfieldAdd( msgs->info->have, ui32 );
|
|
updatePeerProgress( msgs );
|
|
tr_rcTransferred( msgs->torrent->swarmspeed, msgs->torrent->info.pieceSize );
|
|
break;
|
|
|
|
case BT_BITFIELD: {
|
|
const int clientIsSeed = tr_torrentIsSeed( msgs->torrent );
|
|
dbgmsg( msgs, "got a bitfield" );
|
|
tr_peerIoReadBytes( msgs->io, inbuf, msgs->info->have->bits, msglen );
|
|
updatePeerProgress( msgs );
|
|
tr_peerMsgsSetChoke( msgs, !clientIsSeed || (msgs->info->progress<1.0) );
|
|
fireNeedReq( msgs );
|
|
break;
|
|
}
|
|
|
|
case BT_REQUEST: {
|
|
struct peer_request req;
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.index );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.offset );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.length );
|
|
dbgmsg( msgs, "got Request: %u:%u->%u", req.index, req.offset, req.length );
|
|
peerMadeRequest( msgs, &req );
|
|
break;
|
|
}
|
|
|
|
case BT_CANCEL: {
|
|
struct peer_request req;
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.index );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.offset );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.length );
|
|
dbgmsg( msgs, "got a Cancel %u:%u->%u", req.index, req.offset, req.length );
|
|
tr_free( tr_list_remove( &msgs->peerAskedForFast, &req, compareRequest ) );
|
|
tr_free( tr_list_remove( &msgs->peerAskedFor, &req, compareRequest ) );
|
|
break;
|
|
}
|
|
|
|
case BT_PIECE: {
|
|
dbgmsg( msgs, "got a Piece!" );
|
|
assert( msgs->blockToUs.length == 0 );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &msgs->blockToUs.index );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &msgs->blockToUs.offset );
|
|
msgs->blockToUs.length = msglen - 8;
|
|
assert( EVBUFFER_LENGTH(msgs->inBlock) == 0 );
|
|
msgs->state = msgs->blockToUs.length ? READING_BT_PIECE : AWAITING_BT_LENGTH;
|
|
return READ_AGAIN;
|
|
break;
|
|
}
|
|
|
|
case BT_PORT: {
|
|
dbgmsg( msgs, "Got a BT_PORT" );
|
|
tr_peerIoReadUint16( msgs->io, inbuf, &msgs->info->port );
|
|
break;
|
|
}
|
|
|
|
case BT_SUGGEST: {
|
|
/* FIXME(tiennou) */
|
|
uint32_t index;
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &index );
|
|
break;
|
|
}
|
|
|
|
case BT_HAVE_ALL:
|
|
dbgmsg( msgs, "Got a BT_HAVE_ALL" );
|
|
tr_bitfieldAddRange( msgs->info->have, 0, msgs->torrent->info.pieceCount );
|
|
updatePeerProgress( msgs );
|
|
break;
|
|
|
|
case BT_HAVE_NONE:
|
|
dbgmsg( msgs, "Got a BT_HAVE_NONE" );
|
|
tr_bitfieldClear( msgs->info->have );
|
|
updatePeerProgress( msgs );
|
|
break;
|
|
|
|
case BT_REJECT: {
|
|
struct peer_request req;
|
|
dbgmsg( msgs, "Got a BT_REJECT" );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.index );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.offset );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &req.length );
|
|
tr_free( tr_list_remove( &msgs->clientAskedFor, &req, compareRequest ) );
|
|
break;
|
|
}
|
|
|
|
case BT_ALLOWED_FAST: {
|
|
dbgmsg( msgs, "Got a BT_ALLOWED_FAST" );
|
|
tr_peerIoReadUint32( msgs->io, inbuf, &ui32 );
|
|
tr_bitfieldAdd( msgs->clientAllowedPieces, ui32 );
|
|
break;
|
|
}
|
|
|
|
case BT_LTEP:
|
|
dbgmsg( msgs, "Got a BT_LTEP" );
|
|
parseLtep( msgs, msglen, inbuf );
|
|
break;
|
|
|
|
default:
|
|
dbgmsg( msgs, "peer sent us an UNKNOWN: %d", (int)id );
|
|
tr_peerIoDrain( msgs->io, inbuf, msglen );
|
|
break;
|
|
}
|
|
|
|
msgs->incomingMessageLength = -1;
|
|
msgs->state = AWAITING_BT_LENGTH;
|
|
return READ_AGAIN;
|
|
}
|
|
|
|
static void
|
|
clientGotBytes( tr_peermsgs * msgs, uint32_t byteCount )
|
|
{
|
|
tr_torrent * tor = msgs->torrent;
|
|
tor->activityDate = tr_date( );
|
|
tor->downloadedCur += byteCount;
|
|
msgs->info->pieceDataActivityDate = time( NULL );
|
|
tr_rcTransferred( msgs->info->rcToClient, byteCount );
|
|
tr_rcTransferred( tor->download, byteCount );
|
|
tr_rcTransferred( tor->handle->download, byteCount );
|
|
}
|
|
|
|
static void
|
|
peerGotBytes( tr_peermsgs * msgs, uint32_t byteCount )
|
|
{
|
|
tr_torrent * tor = msgs->torrent;
|
|
tor->activityDate = tr_date( );
|
|
tor->uploadedCur += byteCount;
|
|
msgs->info->pieceDataActivityDate = time( NULL );
|
|
tr_rcTransferred( msgs->info->rcToPeer, byteCount );
|
|
tr_rcTransferred( tor->upload, byteCount );
|
|
tr_rcTransferred( tor->handle->upload, byteCount );
|
|
}
|
|
|
|
static int
|
|
canDownload( const tr_peermsgs * msgs )
|
|
{
|
|
tr_torrent * tor = msgs->torrent;
|
|
|
|
if( tor->downloadLimitMode == TR_SPEEDLIMIT_GLOBAL )
|
|
return !tor->handle->useDownloadLimit || tr_rcCanTransfer( tor->handle->download );
|
|
|
|
if( tor->downloadLimitMode == TR_SPEEDLIMIT_SINGLE )
|
|
return tr_rcCanTransfer( tor->download );
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static void
|
|
reassignBytesToCorrupt( tr_peermsgs * msgs, uint32_t byteCount )
|
|
{
|
|
tr_torrent * tor = msgs->torrent;
|
|
|
|
/* increment the `corrupt' field */
|
|
tor->corruptCur += byteCount;
|
|
|
|
/* decrement the `downloaded' field */
|
|
if( tor->downloadedCur >= byteCount )
|
|
tor->downloadedCur -= byteCount;
|
|
else
|
|
tor->downloadedCur = 0;
|
|
}
|
|
|
|
|
|
static void
|
|
gotBadPiece( tr_peermsgs * msgs, uint32_t pieceIndex )
|
|
{
|
|
const uint32_t byteCount = tr_torPieceCountBytes( msgs->torrent, (int)pieceIndex );
|
|
reassignBytesToCorrupt( msgs, byteCount );
|
|
}
|
|
|
|
static void
|
|
gotUnwantedBlock( tr_peermsgs * msgs,
|
|
uint32_t index UNUSED,
|
|
uint32_t offset UNUSED,
|
|
uint32_t length )
|
|
{
|
|
reassignBytesToCorrupt( msgs, length );
|
|
}
|
|
|
|
static void
|
|
addUsToBlamefield( tr_peermsgs * msgs, uint32_t index )
|
|
{
|
|
if( !msgs->info->blame )
|
|
msgs->info->blame = tr_bitfieldNew( msgs->torrent->info.pieceCount );
|
|
tr_bitfieldAdd( msgs->info->blame, index );
|
|
}
|
|
|
|
static void
|
|
gotBlock( tr_peermsgs * msgs,
|
|
struct evbuffer * inbuf,
|
|
uint32_t index,
|
|
uint32_t offset,
|
|
uint32_t length )
|
|
{
|
|
tr_torrent * tor = msgs->torrent;
|
|
const int block = _tr_block( tor, index, offset );
|
|
struct peer_request key, *req;
|
|
|
|
/**
|
|
*** Remove the block from our `we asked for this' list
|
|
**/
|
|
|
|
key.index = index;
|
|
key.offset = offset;
|
|
key.length = length;
|
|
req = (struct peer_request*) tr_list_remove( &msgs->clientAskedFor, &key,
|
|
compareRequest );
|
|
if( req == NULL ) {
|
|
gotUnwantedBlock( msgs, index, offset, length );
|
|
dbgmsg( msgs, "we didn't ask for this message..." );
|
|
return;
|
|
}
|
|
dbgmsg( msgs, "got block %u:%u->%u (turnaround time %d secs)",
|
|
req->index, req->offset, req->length,
|
|
(int)(time(NULL) - req->time_requested) );
|
|
tr_free( req );
|
|
dbgmsg( msgs, "peer has %d more blocks we've asked for",
|
|
tr_list_size(msgs->clientAskedFor));
|
|
|
|
/**
|
|
*** Error checks
|
|
**/
|
|
|
|
if( tr_cpBlockIsComplete( tor->completion, block ) ) {
|
|
dbgmsg( msgs, "have this block already..." );
|
|
tr_dbg( "have this block already..." );
|
|
gotUnwantedBlock( msgs, index, offset, length );
|
|
return;
|
|
}
|
|
|
|
if( (int)length != tr_torBlockCountBytes( tor, block ) ) {
|
|
dbgmsg( msgs, "block is the wrong length..." );
|
|
tr_dbg( "block is the wrong length..." );
|
|
gotUnwantedBlock( msgs, index, offset, length );
|
|
return;
|
|
}
|
|
|
|
/**
|
|
*** Write the block
|
|
**/
|
|
|
|
if( tr_ioWrite( tor, index, offset, length, EVBUFFER_DATA( inbuf )))
|
|
return;
|
|
|
|
#warning this sanity check is here to help track down the excess corrupt data bug, but is expensive and should be removed before the next release
|
|
{
|
|
uint8_t * tmp = tr_new( uint8_t, length );
|
|
const int val = tr_ioRead( tor, index, offset, length, tmp );
|
|
assert( !val );
|
|
assert( !memcmp( tmp, EVBUFFER_DATA(inbuf), length ) );
|
|
tr_free( tmp );
|
|
}
|
|
|
|
tr_cpBlockAdd( tor->completion, block );
|
|
|
|
addUsToBlamefield( msgs, index );
|
|
|
|
fireGotBlock( msgs, index, offset, length );
|
|
|
|
/**
|
|
*** Handle if this was the last block in the piece
|
|
**/
|
|
|
|
if( tr_cpPieceIsComplete( tor->completion, index ) )
|
|
{
|
|
if( tr_ioHash( tor, index ) )
|
|
{
|
|
gotBadPiece( msgs, index );
|
|
return;
|
|
}
|
|
|
|
fireClientHave( msgs, index );
|
|
}
|
|
}
|
|
|
|
|
|
static ReadState
|
|
readBtPiece( tr_peermsgs * msgs, struct evbuffer * inbuf )
|
|
{
|
|
uint32_t inlen;
|
|
uint8_t * tmp;
|
|
|
|
assert( msgs != NULL );
|
|
assert( msgs->blockToUs.length > 0 );
|
|
assert( inbuf != NULL );
|
|
assert( EVBUFFER_LENGTH( inbuf ) > 0 );
|
|
|
|
/* read from the inbuf into our block buffer */
|
|
inlen = MIN( EVBUFFER_LENGTH(inbuf), msgs->blockToUs.length );
|
|
tmp = tr_new( uint8_t, inlen );
|
|
tr_peerIoReadBytes( msgs->io, inbuf, tmp, inlen );
|
|
evbuffer_add( msgs->inBlock, tmp, inlen );
|
|
|
|
/* update our tables accordingly */
|
|
assert( inlen >= msgs->blockToUs.length );
|
|
msgs->blockToUs.length -= inlen;
|
|
msgs->info->peerSentPieceDataAt = time( NULL );
|
|
clientGotBytes( msgs, inlen );
|
|
|
|
/* if this was the entire block, save it */
|
|
if( !msgs->blockToUs.length )
|
|
{
|
|
dbgmsg( msgs, "got block %u:%u", msgs->blockToUs.index, msgs->blockToUs.offset );
|
|
assert( (int)EVBUFFER_LENGTH( msgs->inBlock ) == tr_torBlockCountBytes( msgs->torrent, _tr_block(msgs->torrent,msgs->blockToUs.index, msgs->blockToUs.offset) ) );
|
|
gotBlock( msgs, msgs->inBlock,
|
|
msgs->blockToUs.index,
|
|
msgs->blockToUs.offset,
|
|
EVBUFFER_LENGTH( msgs->inBlock ) );
|
|
evbuffer_drain( msgs->inBlock, ~0 );
|
|
msgs->state = AWAITING_BT_LENGTH;
|
|
}
|
|
|
|
/* cleanup */
|
|
tr_free( tmp );
|
|
return READ_AGAIN;
|
|
}
|
|
|
|
static ReadState
|
|
canRead( struct bufferevent * evin, void * vmsgs )
|
|
{
|
|
ReadState ret;
|
|
tr_peermsgs * msgs = (tr_peermsgs *) vmsgs;
|
|
struct evbuffer * inbuf = EVBUFFER_INPUT ( evin );
|
|
|
|
if( !canDownload( msgs ) )
|
|
{
|
|
msgs->notListening = 1;
|
|
tr_peerIoSetIOMode ( msgs->io, 0, EV_READ );
|
|
ret = READ_DONE;
|
|
}
|
|
else switch( msgs->state )
|
|
{
|
|
case AWAITING_BT_LENGTH: ret = readBtLength ( msgs, inbuf ); break;
|
|
case AWAITING_BT_MESSAGE: ret = readBtMessage ( msgs, inbuf ); break;
|
|
case READING_BT_PIECE: ret = readBtPiece ( msgs, inbuf ); break;
|
|
default: assert( 0 );
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
static void
|
|
sendKeepalive( tr_peermsgs * msgs )
|
|
{
|
|
dbgmsg( msgs, "sending a keepalive message" );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, 0 );
|
|
}
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
static int
|
|
canWrite( const tr_peermsgs * msgs )
|
|
{
|
|
/* don't let our outbuffer get too large */
|
|
if( tr_peerIoWriteBytesWaiting( msgs->io ) > 4096 )
|
|
return FALSE;
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static size_t
|
|
getUploadMax( const tr_peermsgs * msgs )
|
|
{
|
|
static const size_t maxval = ~0;
|
|
const tr_torrent * tor = msgs->torrent;
|
|
|
|
if( !canWrite( msgs ) )
|
|
return 0;
|
|
|
|
if( tor->uploadLimitMode == TR_SPEEDLIMIT_GLOBAL )
|
|
return tor->handle->useUploadLimit ? tr_rcBytesLeft( tor->handle->upload ) : maxval;
|
|
|
|
if( tor->uploadLimitMode == TR_SPEEDLIMIT_SINGLE )
|
|
return tr_rcBytesLeft( tor->upload );
|
|
|
|
return maxval;
|
|
}
|
|
|
|
static int
|
|
ratePulse( void * vmsgs )
|
|
{
|
|
tr_peermsgs * msgs = (tr_peermsgs *) vmsgs;
|
|
msgs->info->rateToClient = tr_rcRate( msgs->info->rcToClient );
|
|
msgs->info->rateToPeer = tr_rcRate( msgs->info->rcToPeer );
|
|
msgs->maxActiveRequests = MIN( 8 + (int)(msgs->info->rateToClient/10), 100 );
|
|
msgs->minActiveRequests = msgs->maxActiveRequests / 2;
|
|
return TRUE;
|
|
}
|
|
|
|
static struct peer_request*
|
|
popNextRequest( tr_peermsgs * msgs )
|
|
{
|
|
struct peer_request * ret;
|
|
ret = tr_list_pop_front( &msgs->peerAskedForFast );
|
|
if( !ret )
|
|
ret = tr_list_pop_front( &msgs->peerAskedFor);
|
|
return ret;
|
|
}
|
|
|
|
static int
|
|
pulse( void * vmsgs )
|
|
{
|
|
const time_t now = time( NULL );
|
|
tr_peermsgs * msgs = vmsgs;
|
|
struct peer_request * r;
|
|
size_t len;
|
|
|
|
/* if we froze out a downloaded block because of speed limits,
|
|
start listening to the peer again */
|
|
if( msgs->notListening && canDownload( msgs ) )
|
|
{
|
|
msgs->notListening = 0;
|
|
tr_peerIoSetIOMode ( msgs->io, EV_READ, 0 );
|
|
}
|
|
|
|
pumpRequestQueue( msgs );
|
|
|
|
if( !canWrite( msgs ) )
|
|
{
|
|
}
|
|
else if(( len = EVBUFFER_LENGTH( msgs->outBlock ) ))
|
|
{
|
|
const size_t uploadMax = getUploadMax( msgs );
|
|
const size_t outlen = MIN( len, uploadMax );
|
|
tr_peerIoWrite( msgs->io, EVBUFFER_DATA( msgs->outBlock ), outlen );
|
|
evbuffer_drain( msgs->outBlock, outlen );
|
|
msgs->clientSentAnythingAt = now;
|
|
peerGotBytes( msgs, outlen );
|
|
len -= outlen;
|
|
dbgmsg( msgs, "wrote %d bytes; %d left in block", (int)outlen, (int)len );
|
|
fflush( stdout );
|
|
}
|
|
else if(( len = EVBUFFER_LENGTH( msgs->outMessages ) ))
|
|
{
|
|
tr_peerIoWriteBuf( msgs->io, msgs->outMessages );
|
|
msgs->clientSentAnythingAt = now;
|
|
}
|
|
else if( ( now - msgs->clientSentAnythingAt ) > KEEPALIVE_INTERVAL_SECS )
|
|
{
|
|
sendKeepalive( msgs );
|
|
}
|
|
|
|
if( !EVBUFFER_LENGTH( msgs->outBlock )
|
|
&& (( r = popNextRequest( msgs )))
|
|
&& requestIsValid( msgs, r )
|
|
&& tr_cpPieceIsComplete( msgs->torrent->completion, r->index ) )
|
|
{
|
|
uint8_t * buf = tr_new( uint8_t, r->length );
|
|
|
|
if( !tr_ioRead( msgs->torrent, r->index, r->offset, r->length, buf ) )
|
|
{
|
|
tr_peerIo * io = msgs->io;
|
|
struct evbuffer * out = msgs->outBlock;
|
|
|
|
dbgmsg( msgs, "sending block %u:%u->%u", r->index, r->offset, r->length );
|
|
tr_peerIoWriteUint32( io, out, sizeof(uint8_t) + 2*sizeof(uint32_t) + r->length );
|
|
tr_peerIoWriteUint8 ( io, out, BT_PIECE );
|
|
tr_peerIoWriteUint32( io, out, r->index );
|
|
tr_peerIoWriteUint32( io, out, r->offset );
|
|
tr_peerIoWriteBytes ( io, out, buf, r->length );
|
|
}
|
|
|
|
tr_free( buf );
|
|
tr_free( r );
|
|
|
|
pulse( msgs ); /* start sending it right away */
|
|
}
|
|
|
|
return TRUE; /* loop forever */
|
|
}
|
|
|
|
static void
|
|
didWrite( struct bufferevent * evin UNUSED, void * vmsgs )
|
|
{
|
|
pulse( vmsgs );
|
|
}
|
|
|
|
static void
|
|
gotError( struct bufferevent * evbuf UNUSED, short what, void * vmsgs )
|
|
{
|
|
dbgmsg( vmsgs, "libevent got an error! what=%d, errno=%d (%s)",
|
|
(int)what, errno, strerror(errno) );
|
|
fireGotError( vmsgs );
|
|
}
|
|
|
|
static void
|
|
sendBitfield( tr_peermsgs * msgs )
|
|
{
|
|
const tr_bitfield * bitfield = tr_cpPieceBitfield( msgs->torrent->completion );
|
|
struct evbuffer * out = msgs->outMessages;
|
|
|
|
dbgmsg( msgs, "sending peer a bitfield message" );
|
|
tr_peerIoWriteUint32( msgs->io, out, sizeof(uint8_t) + bitfield->len );
|
|
tr_peerIoWriteUint8 ( msgs->io, out, BT_BITFIELD );
|
|
tr_peerIoWriteBytes ( msgs->io, out, bitfield->bits, bitfield->len );
|
|
}
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
/* some peers give us error messages if we send
|
|
more than this many peers in a single pex message */
|
|
#define MAX_PEX_DIFFS 200
|
|
|
|
typedef struct
|
|
{
|
|
tr_pex * added;
|
|
tr_pex * dropped;
|
|
tr_pex * elements;
|
|
int addedCount;
|
|
int droppedCount;
|
|
int elementCount;
|
|
int diffCount;
|
|
}
|
|
PexDiffs;
|
|
|
|
static void
|
|
pexAddedCb( void * vpex, void * userData )
|
|
{
|
|
PexDiffs * diffs = (PexDiffs *) userData;
|
|
tr_pex * pex = (tr_pex *) vpex;
|
|
if( diffs->diffCount < MAX_PEX_DIFFS )
|
|
{
|
|
diffs->diffCount++;
|
|
diffs->added[diffs->addedCount++] = *pex;
|
|
diffs->elements[diffs->elementCount++] = *pex;
|
|
}
|
|
}
|
|
|
|
static void
|
|
pexRemovedCb( void * vpex, void * userData )
|
|
{
|
|
PexDiffs * diffs = (PexDiffs *) userData;
|
|
tr_pex * pex = (tr_pex *) vpex;
|
|
if( diffs->diffCount < MAX_PEX_DIFFS )
|
|
{
|
|
diffs->diffCount++;
|
|
diffs->dropped[diffs->droppedCount++] = *pex;
|
|
}
|
|
}
|
|
|
|
static void
|
|
pexElementCb( void * vpex, void * userData )
|
|
{
|
|
PexDiffs * diffs = (PexDiffs *) userData;
|
|
tr_pex * pex = (tr_pex *) vpex;
|
|
if( diffs->diffCount < MAX_PEX_DIFFS )
|
|
{
|
|
diffs->diffCount++;
|
|
diffs->elements[diffs->elementCount++] = *pex;
|
|
}
|
|
}
|
|
|
|
static void
|
|
sendPex( tr_peermsgs * msgs )
|
|
{
|
|
if( msgs->peerSupportsPex && tr_torrentIsPexEnabled( msgs->torrent ) )
|
|
{
|
|
int i;
|
|
tr_pex * newPex = NULL;
|
|
const int newCount = tr_peerMgrGetPeers( msgs->handle->peerMgr, msgs->torrent->info.hash, &newPex );
|
|
PexDiffs diffs;
|
|
benc_val_t val, *added, *dropped, *flags;
|
|
uint8_t *tmp, *walk;
|
|
char * benc;
|
|
int bencLen;
|
|
|
|
/* build the diffs */
|
|
diffs.added = tr_new( tr_pex, newCount );
|
|
diffs.addedCount = 0;
|
|
diffs.dropped = tr_new( tr_pex, msgs->pexCount );
|
|
diffs.droppedCount = 0;
|
|
diffs.elements = tr_new( tr_pex, newCount + msgs->pexCount );
|
|
diffs.elementCount = 0;
|
|
diffs.diffCount = 0;
|
|
tr_set_compare( msgs->pex, msgs->pexCount,
|
|
newPex, newCount,
|
|
tr_pexCompare, sizeof(tr_pex),
|
|
pexRemovedCb, pexAddedCb, pexElementCb, &diffs );
|
|
dbgmsg( msgs, "pex: old peer count %d, new peer count %d, added %d, removed %d", msgs->pexCount, newCount, diffs.addedCount, diffs.droppedCount );
|
|
|
|
/* update peer */
|
|
tr_free( msgs->pex );
|
|
msgs->pex = diffs.elements;
|
|
msgs->pexCount = diffs.elementCount;
|
|
|
|
/* build the pex payload */
|
|
tr_bencInit( &val, TYPE_DICT );
|
|
tr_bencDictReserve( &val, 3 );
|
|
|
|
/* "added" */
|
|
added = tr_bencDictAdd( &val, "added" );
|
|
tmp = walk = tr_new( uint8_t, diffs.addedCount * 6 );
|
|
for( i=0; i<diffs.addedCount; ++i ) {
|
|
memcpy( walk, &diffs.added[i].in_addr, 4 ); walk += 4;
|
|
memcpy( walk, &diffs.added[i].port, 2 ); walk += 2;
|
|
}
|
|
assert( ( walk - tmp ) == diffs.addedCount * 6 );
|
|
tr_bencInitStr( added, tmp, walk-tmp, FALSE );
|
|
|
|
/* "added.f" */
|
|
flags = tr_bencDictAdd( &val, "added.f" );
|
|
tmp = walk = tr_new( uint8_t, diffs.addedCount );
|
|
for( i=0; i<diffs.addedCount; ++i )
|
|
*walk++ = diffs.added[i].flags;
|
|
assert( ( walk - tmp ) == diffs.addedCount );
|
|
tr_bencInitStr( flags, tmp, walk-tmp, FALSE );
|
|
|
|
/* "dropped" */
|
|
dropped = tr_bencDictAdd( &val, "dropped" );
|
|
tmp = walk = tr_new( uint8_t, diffs.droppedCount * 6 );
|
|
for( i=0; i<diffs.droppedCount; ++i ) {
|
|
memcpy( walk, &diffs.dropped[i].in_addr, 4 ); walk += 4;
|
|
memcpy( walk, &diffs.dropped[i].port, 2 ); walk += 2;
|
|
}
|
|
assert( ( walk - tmp ) == diffs.droppedCount * 6 );
|
|
tr_bencInitStr( dropped, tmp, walk-tmp, FALSE );
|
|
|
|
/* write the pex message */
|
|
benc = tr_bencSaveMalloc( &val, &bencLen );
|
|
tr_peerIoWriteUint32( msgs->io, msgs->outMessages, 2*sizeof(uint8_t) + bencLen );
|
|
tr_peerIoWriteUint8 ( msgs->io, msgs->outMessages, BT_LTEP );
|
|
tr_peerIoWriteUint8 ( msgs->io, msgs->outMessages, OUR_LTEP_PEX );
|
|
tr_peerIoWriteBytes ( msgs->io, msgs->outMessages, benc, bencLen );
|
|
|
|
/* cleanup */
|
|
tr_free( benc );
|
|
tr_bencFree( &val );
|
|
tr_free( diffs.added );
|
|
tr_free( diffs.dropped );
|
|
tr_free( newPex );
|
|
|
|
msgs->clientSentPexAt = time( NULL );
|
|
}
|
|
}
|
|
|
|
static int
|
|
pexPulse( void * vpeer )
|
|
{
|
|
sendPex( vpeer );
|
|
return TRUE;
|
|
}
|
|
|
|
/**
|
|
***
|
|
**/
|
|
|
|
tr_peermsgs*
|
|
tr_peerMsgsNew( struct tr_torrent * torrent,
|
|
struct tr_peer * info,
|
|
tr_delivery_func func,
|
|
void * userData,
|
|
tr_publisher_tag * setme )
|
|
{
|
|
tr_peermsgs * m;
|
|
|
|
assert( info != NULL );
|
|
assert( info->io != NULL );
|
|
|
|
m = tr_new0( tr_peermsgs, 1 );
|
|
m->publisher = tr_publisherNew( );
|
|
m->info = info;
|
|
m->handle = torrent->handle;
|
|
m->torrent = torrent;
|
|
m->io = info->io;
|
|
m->info->clientIsChoked = 1;
|
|
m->info->peerIsChoked = 1;
|
|
m->info->clientIsInterested = 0;
|
|
m->info->peerIsInterested = 0;
|
|
m->info->have = tr_bitfieldNew( torrent->info.pieceCount );
|
|
m->pulseTimer = tr_timerNew( m->handle, pulse, m, PEER_PULSE_INTERVAL );
|
|
m->rateTimer = tr_timerNew( m->handle, ratePulse, m, RATE_PULSE_INTERVAL );
|
|
m->pexTimer = tr_timerNew( m->handle, pexPulse, m, PEX_INTERVAL );
|
|
m->outMessages = evbuffer_new( );
|
|
m->outBlock = evbuffer_new( );
|
|
m->inBlock = evbuffer_new( );
|
|
m->peerAllowedPieces = NULL;
|
|
m->clientAllowedPieces = NULL;
|
|
*setme = tr_publisherSubscribe( m->publisher, func, userData );
|
|
|
|
if ( tr_peerIoSupportsFEXT( m->io ) )
|
|
{
|
|
/* This peer is fastpeer-enabled, generate its allowed set
|
|
* (before registering our callbacks) */
|
|
if ( !m->peerAllowedPieces ) {
|
|
const struct in_addr *peerAddr = tr_peerIoGetAddress( m->io, NULL );
|
|
|
|
m->peerAllowedPieces = tr_peerMgrGenerateAllowedSet( MAX_ALLOWED_SET_COUNT,
|
|
m->torrent->info.pieceCount,
|
|
m->torrent->info.hash,
|
|
peerAddr );
|
|
}
|
|
m->clientAllowedPieces = tr_bitfieldNew( m->torrent->info.pieceCount );
|
|
}
|
|
|
|
tr_peerIoSetTimeoutSecs( m->io, 150 ); /* error if we don't read or write for 2.5 minutes */
|
|
tr_peerIoSetIOFuncs( m->io, canRead, didWrite, gotError, m );
|
|
tr_peerIoSetIOMode( m->io, EV_READ|EV_WRITE, 0 );
|
|
ratePulse( m );
|
|
|
|
if ( tr_peerIoSupportsLTEP( m->io ) )
|
|
sendLtepHandshake( m );
|
|
|
|
if ( !tr_peerIoSupportsFEXT( m->io ) )
|
|
sendBitfield( m );
|
|
else {
|
|
/* This peer is fastpeer-enabled, send it have-all or have-none if appropriate */
|
|
uint32_t peerProgress;
|
|
float completion = tr_cpPercentComplete( m->torrent->completion );
|
|
if ( completion == 0.0f ) {
|
|
sendFastHave( m, 0 );
|
|
} else if ( completion == 1.0f ) {
|
|
sendFastHave( m, 1 );
|
|
} else {
|
|
sendBitfield( m );
|
|
}
|
|
peerProgress = m->torrent->info.pieceCount * m->info->progress;
|
|
|
|
if ( peerProgress < MAX_ALLOWED_SET_COUNT )
|
|
sendFastAllowedSet( m );
|
|
}
|
|
|
|
return m;
|
|
}
|
|
|
|
void
|
|
tr_peerMsgsFree( tr_peermsgs* msgs )
|
|
{
|
|
if( msgs != NULL )
|
|
{
|
|
tr_timerFree( &msgs->pulseTimer );
|
|
tr_timerFree( &msgs->rateTimer );
|
|
tr_timerFree( &msgs->pexTimer );
|
|
tr_publisherFree( &msgs->publisher );
|
|
tr_list_free( &msgs->clientWillAskFor, tr_free );
|
|
tr_list_free( &msgs->clientAskedFor, tr_free );
|
|
tr_list_free( &msgs->peerAskedForFast, tr_free );
|
|
tr_list_free( &msgs->peerAskedFor, tr_free );
|
|
evbuffer_free( msgs->outMessages );
|
|
evbuffer_free( msgs->outBlock );
|
|
evbuffer_free( msgs->inBlock );
|
|
tr_free( msgs->pex );
|
|
msgs->pexCount = 0;
|
|
tr_free( msgs );
|
|
}
|
|
}
|
|
|
|
tr_publisher_tag
|
|
tr_peerMsgsSubscribe( tr_peermsgs * peer,
|
|
tr_delivery_func func,
|
|
void * userData )
|
|
{
|
|
return tr_publisherSubscribe( peer->publisher, func, userData );
|
|
}
|
|
|
|
void
|
|
tr_peerMsgsUnsubscribe( tr_peermsgs * peer,
|
|
tr_publisher_tag tag )
|
|
{
|
|
tr_publisherUnsubscribe( peer->publisher, tag );
|
|
}
|
|
|
|
int
|
|
tr_peerMsgIsPieceFastAllowed( const tr_peermsgs * peer,
|
|
uint32_t index )
|
|
{
|
|
return tr_bitfieldHas( peer->clientAllowedPieces, index );
|
|
}
|