R5Reloaded/r5sdk
1
0
Fork 0
mirror of https://github.com/Mauler125/r5sdk.git synced 2025-02-09 19:15:03 +01:00
Code Issues Packages Projects Releases Wiki Activity

CNetChan::FlowNewPacket(): Add a comment and also check the other loop

Browse Source 
Added 2 comments that speaks for them self's, and 2 additional checks for the second loop as I had triggered several cases where it would loop indefinitely still, as we break out of the parallel loop and continue into the old school loop for the remainder of the outSeqNr - currentIndex frames. This should fix all problems related to this function. After numerous of tests with varying arbitrary outSeqNr values that were vastly larger than the current packet flow index, the bug was no longer triggered. A similar test had also be performed were multiple clients would do this at the same time, and even then; the server frame time(s) remained identical and the whole thing never hung.
This commit is contained in:
Amos 2023-08-18 21:24:35 +02:00
parent 26da5fd503
commit fe1cc93cff
1 changed files with 30 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
r5dev/engine/net_chan.cpp
View File

@ -81,7 +81,7 @@ void CNetChan::_FlowNewPacket(CNetChan* pChan, int flow, int outSeqNr, int inSeq
int v9; // r14d
int v12; // r12d
int currentindex; // eax
int v16; // r15d
int nextIndex; // r15d
int v17; // r8d
int v18; // ebp
unsigned int v19; // eax
@ -124,26 +124,32 @@ void CNetChan::_FlowNewPacket(CNetChan* pChan, int flow, int outSeqNr, int inSeq
currentindex = pFlow->currentindex;
if (outSeqNr > currentindex)
{
v16 = currentindex + 1;
nextIndex = currentindex + 1;
if (currentindex + 1 <= outSeqNr)
{
v17 = outSeqNr - v16;
// This variable makes sure the loops below do not execute more
// than NET_FRAMES_BACKUP times. This has to be done as the
// headers and frame arrays in the netflow_t structure is as
// large as NET_FRAMES_BACKUP. Any execution past it is futile
// and only wastes CPU time. Sending an outSeqNr that is higher
// than the current index by something like a million or more will
// hang the engine for several milliseconds to several seconds.
int numPacketFrames = 0;
v17 = outSeqNr - nextIndex;
if (v17 + 1 >= 4)
{
v18 = nChoked + nDropped;
v19 = ((unsigned int)(v12 - v16 - 3) >> 2) + 1;
v20 = v16 + 2;
v19 = ((unsigned int)(v12 - nextIndex - 3) >> 2) + 1;
v20 = nextIndex + 2;
v21 = v17 - 2;
v22 = v19;
time = (float)*g_pNetTime;
v16 += 4 * v19;
int numPacketFrames = 0;
nextIndex += 4 * v19;
do
{
numPacketFrames += 4;
v24 = (v20 - 2) & NET_FRAMES_MASK;
v25 = v24;
pFlow->frame_headers[v25].time = time;
@ -208,6 +214,10 @@ void CNetChan::_FlowNewPacket(CNetChan* pChan, int flow, int outSeqNr, int inSeq
else
pFrameHeader->choked = 1;
}
// Incremented by four since this loop does four frames
// per iteration.
numPacketFrames += 4;
v21 -= 4;
v20 += 4;
--v22;
@ -216,11 +226,15 @@ void CNetChan::_FlowNewPacket(CNetChan* pChan, int flow, int outSeqNr, int inSeq
v8 = flow;
v9 = inSeqNr;
}
if (v16 <= v12)
// Check if we did not reach NET_FRAMES_BACKUP, else we will
// execute the 129'th iteration as well. Also check if the next
// index doesn't exceed the outSeqNr.
if (numPacketFrames < NET_FRAMES_BACKUP && nextIndex <= v12)
{
v30 = v12 - v16;
v31 = v16;
v33 = v12 - v16 + 1;
v30 = v12 - nextIndex;
v31 = nextIndex;
v33 = v12 - nextIndex + 1;
do
{
pFrame = &pFlow->frames[v31 & NET_FRAMES_MASK];
@ -243,7 +257,8 @@ void CNetChan::_FlowNewPacket(CNetChan* pChan, int flow, int outSeqNr, int inSeq
--v30;
++v31;
--v33;
} while (v33);
++numPacketFrames;
} while (v33 && numPacketFrames < NET_FRAMES_BACKUP);
v9 = inSeqNr;
}
}