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
r5sdk/r5dev/datacache/mdlcache.cpp

322 lines
11 KiB
C++
Raw Normal View History

Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
//=====================================================================================//
//
// model loading and caching
//
// $NoKeywords: $
//=====================================================================================//
#include "core/stdafx.h"
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
#include "tier0/threadtools.h"
#include "tier1/cvar.h"
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
#include "datacache/mdlcache.h"
#include "datacache/imdlcache.h"
#include "engine/sys_utils.h"
#include "rtech/rtech_utils.h"
#include "public/include/studio.h"
IClientEntityList and CUtlDict. * IClientEntityList is the full virtual function table. * CUtlDict "implementation" is temporary.
2022-05-01 23:03:20 +02:00
#include "tier1/utldict.h"
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
//-----------------------------------------------------------------------------
// Purpose: finds an MDL
// Input : *this -
// handle -
// *a3 -
// Output : a pointer to the studiohdr_t object
//-----------------------------------------------------------------------------
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
studiohdr_t* CMDLCache::FindMDL(CMDLCache* cache, MDLHandle_t handle, void* a3)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
studiodata_t* pStudioData; // rbx
void* pMDLCache; // rax
studiohdr_t* result; // rax
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
EnterCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
auto mdlDict = CUtlDict<studiodata_t*, MDLHandle_t>(m_MDLDict.Deref().GetPtr());
pStudioData = mdlDict.Find(static_cast<int64>(handle));
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
LeaveCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
if (!g_pMDLFallback->m_hErrorMDL || !g_pMDLFallback->m_hEmptyMDL)
{
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
studiohdr_t* pStudioHDR = **reinterpret_cast<studiohdr_t***>(pStudioData);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
string svStudio = ConvertToUnixPath(string(pStudioHDR->name));
if (strcmp(svStudio.c_str(), ERROR_MODEL) == 0)
{
g_pMDLFallback->m_pErrorHDR = pStudioHDR;
g_pMDLFallback->m_hErrorMDL = handle;
}
if (strcmp(svStudio.c_str(), EMPTY_MODEL) == 0)
{
g_pMDLFallback->m_pEmptyHDR = pStudioHDR;
g_pMDLFallback->m_hEmptyMDL = handle;
}
}
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
if (!pStudioData)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
if (!g_pMDLFallback->m_hErrorMDL)
Error(eDLL_T::ENGINE, "Model with handle \"%hu\" not found and \"%s\" couldn't be loaded.\n", handle, ERROR_MODEL);
return g_pMDLFallback->m_pErrorHDR;
}
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
if ((pStudioData->m_nFlags & STUDIOHDR_FLAGS_NEEDS_DEFERRED_ADDITIVE | STUDIOHDR_FLAGS_OBSOLETE) != 0)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
pMDLCache = *reinterpret_cast<void**>(pStudioData);
if (pStudioData->m_MDLCache)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
if (a3)
{
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
FindCachedMDL(cache, pStudioData, a3);
pMDLCache = *reinterpret_cast<void**>(pStudioData);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
LABEL_6:
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
result = *reinterpret_cast<studiohdr_t**>(pMDLCache);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
if (result)
return result;
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
return FindUncachedMDL(cache, handle, pStudioData, a3);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
pMDLCache = pStudioData->Unk0;
if (pMDLCache)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
goto LABEL_6;
}
Improve CMDLCache::FindMDL
2022-05-05 02:54:17 +02:00
return FindUncachedMDL(cache, handle, pStudioData, a3);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
//-----------------------------------------------------------------------------
// Purpose: finds an MDL cached
// Input : *this -
// *a2 -
// *a3 -
//-----------------------------------------------------------------------------
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
void CMDLCache::FindCachedMDL(CMDLCache* cache, void* a2, void* a3)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
__int64 v6; // rax
if (a3)
{
CThreadFastMutex::WaitForLock((CThreadFastMutex*)a2 + 128);
*(_QWORD*)((int64_t)a3 + 2176) = *(_QWORD*)((int64_t)a2 + 88);
v6 = *(_QWORD*)((int64_t)a2 + 88);
if (v6)
*(_QWORD*)(v6 + 2168) = (int64_t)a3;
*(_QWORD*)((int64_t)a2 + 88) = (int64_t)a3;
*(_QWORD*)((int64_t)a3 + 2160) = (int64_t)cache;
*(_WORD*)((int64_t)a3 + 2184) = *(_WORD*)((int64_t)a2 + 20);
CThreadFastMutex::ReleaseWaiter((CThreadFastMutex*)a2 + 128);
}
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
//-----------------------------------------------------------------------------
// Purpose: finds an MDL uncached
// Input : *this -
// handle -
// *a3 -
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
// *a4 -
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
// Output : a pointer to the studiohdr_t object
//-----------------------------------------------------------------------------
studiohdr_t* CMDLCache::FindUncachedMDL(CMDLCache* cache, MDLHandle_t handle, void* a3, void* a4)
{
const char* v8; // rdi
__int64 v9; // rax
int v10; // eax
const char* v11; // r11
int v12; // eax
const char* v13; // r11
int v14; // eax
bool v16; // zf
studiohdr_t* v17; // rdi
studiohdr_t** v18; // rax
See description Fixed rare crash in CMDLCache when buffer itteration returns invalid pointer (not null!) for handle. Made all variables and functions used by BuildPropStaticFrustumCullMap searched dynamically (these should also work on s0, s1 and s2, they are compared but at the moment untested). TODO: Fix deref crash properly (see bsplib.cpp l291).
2022-05-01 05:38:51 +02:00
bool bOldModel {};
bool bInvalidHandle{};
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
CThreadFastMutex::WaitForLock((CThreadFastMutex*)a3 + 0x80);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
EnterCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
void* modelCache = cache->m_pModelCacheSection;
v8 = (const char*)(*(_QWORD*)((int64)modelCache + 24 * static_cast<int64>(handle) + 8));
LeaveCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
See description Fixed rare crash in CMDLCache when buffer itteration returns invalid pointer (not null!) for handle. Made all variables and functions used by BuildPropStaticFrustumCullMap searched dynamically (these should also work on s0, s1 and s2, they are compared but at the moment untested). TODO: Fix deref crash properly (see bsplib.cpp l291).
2022-05-01 05:38:51 +02:00
if (IsBadReadPtrV2((void*)v8))
{
bInvalidHandle = true;
goto LABEL_ERROR;
}
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
v9 = -1i64;
do
++v9;
while (v8[v9]);
if (v9 < 5 ||
(_stricmp(&v8[v9 - 5], ".rmdl") != 0) &&
(_stricmp(&v8[v9 - 5], ".rrig") != 0) &&
(_stricmp(&v8[v9 - 5], ".rpak") != 0))
{
Also lock the upgrade error behind already known bad model handles Reduce console verbosity
2022-04-30 05:06:59 +02:00
bOldModel = true;
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
goto LABEL_ERROR;
}
g_pRTech->StringToGuid(v8);
v16 = *(_QWORD*)a3 == 0i64;
*(_BYTE*)((int64)a3 + 152) = 0;
if (v16)
{
v18 = *(studiohdr_t***)((int64)a3 + 8);
if (v18)
{
v17 = *v18;
}
else
{
Only print out missing models once Don't print duplicated
2022-04-29 22:46:37 +02:00
LABEL_ERROR:
Improved mod RPak loading between changelevel Note: this does not work reliably still. The only way we could make this work 100% reliable would be to fully rebuild '0x140341D40' in the SDK and load our pak files among with the pre-existing g_pakLoadApi->AsyncLoad() call in this function, as this will ensure everything will be ran synchronously. The current approach by taking some JT fifolock wrapper will only work reliably between one level change, unsure why the second and up fail.
2022-05-04 02:25:27 +02:00
if (std::find(g_BadMDLHandles.begin(), g_BadMDLHandles.end(), handle) == g_BadMDLHandles.end())
Only print out missing models once Don't print duplicated
2022-04-29 22:46:37 +02:00
{
See description Fixed rare crash in CMDLCache when buffer itteration returns invalid pointer (not null!) for handle. Made all variables and functions used by BuildPropStaticFrustumCullMap searched dynamically (these should also work on s0, s1 and s2, they are compared but at the moment untested). TODO: Fix deref crash properly (see bsplib.cpp l291).
2022-05-01 05:38:51 +02:00
if (bInvalidHandle)
Error(eDLL_T::ENGINE, "Model with handle \"hu\" not found; replacing with \"%s\".\n", handle, ERROR_MODEL);
else if (bOldModel)
Error(eDLL_T::ENGINE, "Attempted to load old model \"%s\"; replace with rmdl.\n", v8);
else
Also lock the upgrade error behind already known bad model handles Reduce console verbosity
2022-04-30 05:06:59 +02:00
{
if (g_pMDLFallback->m_hErrorMDL)
Error(eDLL_T::ENGINE, "Model \"%s\" not found; replacing with \"%s\".\n", v8, ERROR_MODEL);
else
Error(eDLL_T::ENGINE, "Model \"%s\" not found and \"%s\" couldn't be loaded.\n", v8, ERROR_MODEL);
}
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
Improved mod RPak loading between changelevel Note: this does not work reliably still. The only way we could make this work 100% reliable would be to fully rebuild '0x140341D40' in the SDK and load our pak files among with the pre-existing g_pakLoadApi->AsyncLoad() call in this function, as this will ensure everything will be ran synchronously. The current approach by taking some JT fifolock wrapper will only work reliably between one level change, unsure why the second and up fail.
2022-05-04 02:25:27 +02:00
g_BadMDLHandles.push_back(handle);
Only print out missing models once Don't print duplicated
2022-04-29 22:46:37 +02:00
}
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
v17 = g_pMDLFallback->m_pErrorHDR;
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
old_gather_props->SetValue(true); // mdl/error.rmdl fallback is not supported (yet) in the new GatherProps solution!
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
}
else
{
v_CMDLCache__FindCachedMDL(cache, a3, a4);
Some mdlcache error handling improvements
2022-05-04 12:44:01 +02:00
if ((__int64)*(studiohdr_t**)a3)
{
if ((__int64)*(studiohdr_t**)a3 == 0xDEADFEEDDEADFEED)
v17 = g_pMDLFallback->m_pErrorHDR;
else
v17 = **(studiohdr_t***)a3;
}
else
v17 = g_pMDLFallback->m_pErrorHDR;
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
CThreadFastMutex::ReleaseWaiter((CThreadFastMutex*)a3 + 128);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
return v17;
}
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
//-----------------------------------------------------------------------------
// Purpose: gets the studiohdr from cache pool by handle
// Input : *this -
// handle -
// Output : a pointer to the studiohdr_t object
//-----------------------------------------------------------------------------
studiohdr_t* CMDLCache::GetStudioHDR(CMDLCache* pMDLCache, MDLHandle_t handle)
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
{
__int64 v2; // rbx
__int64 v3; // rbx
__int64 v4; // rdx
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
studiohdr_t* result = nullptr; // rax
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
if (!handle)
{
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
LABEL_ERROR:
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
if (!g_pMDLFallback->m_hErrorMDL)
Error(eDLL_T::ENGINE, "Model with handle \"%hu\" not found and \"%s\" couldn't be loaded.\n", handle, ERROR_MODEL);
return g_pMDLFallback->m_pErrorHDR;
}
v2 = handle;
EnterCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
v3 = *(_QWORD*)(m_MDLDict.Deref().GetPtr() + 24 * v2 + 16);
LeaveCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
Return proper studio hardware for mdl/error.rmdl Now mdl/error.rmdl draws properly
2022-04-29 22:11:35 +02:00
if (*(_QWORD*)(v3))
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
{
Return proper studio hardware for mdl/error.rmdl Now mdl/error.rmdl draws properly
2022-04-29 22:11:35 +02:00
v4 = *(_QWORD*)(*(_QWORD*)(*(_QWORD*)v3 + 8i64) + 24i64);
if (v4)
result = (studiohdr_t*)(v4 + 16);
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
}
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
return result;
}
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
//-----------------------------------------------------------------------------
// Purpose: gets the studio hardware data reference from cache pool by handle
// Input : *this -
// handle -
// Output : a pointer to the CStudioHWDataRef object
//-----------------------------------------------------------------------------
CStudioHWDataRef* CMDLCache::GetStudioHardwareRef(CMDLCache* cache, MDLHandle_t handle)
{
__int64 v2; // rbx
__int64 v3; // rdi
__int64 v4; // rbx
__int64 result; // rax
v2 = handle;
EnterCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
v3 = *(_QWORD*)(m_MDLDict.Deref().GetPtr() + 24 * v2 + 16);
LeaveCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
if (!v3)
{
Return proper studio hardware for mdl/error.rmdl Now mdl/error.rmdl draws properly
2022-04-29 22:11:35 +02:00
if (!g_pMDLFallback->m_hErrorMDL)
{
Error(eDLL_T::ENGINE, "Studio hardware with handle \"%hu\" not found and \"%s\" couldn't be loaded.\n", handle, ERROR_MODEL);
return nullptr;
}
Some mdlcache error handling improvements
2022-05-04 12:44:01 +02:00
v3 = *(_QWORD*)(m_MDLDict.Deref().GetPtr() + 24i64 * g_pMDLFallback->m_hErrorMDL + 16);
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
}
if (*(_QWORD*)v3)
{
Some mdlcache error handling improvements
2022-05-04 12:44:01 +02:00
if (*(_QWORD*)v3 == 0xDEADFEEDDEADFEED)
return nullptr;
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
v4 = *(_QWORD*)(*(_QWORD*)v3 + 8i64);
AcquireSRWLockExclusive(reinterpret_cast<PSRWLOCK>(&*m_MDLLock));
v_CStudioHWDataRef__SetFlags(reinterpret_cast<CStudioHWDataRef*>(v4), 1i64); // !!! DECLARED INLINE IN < S3 !!!
ReleaseSRWLockExclusive(reinterpret_cast<PSRWLOCK>(&*m_MDLLock));
}
if ((*(_BYTE*)(v3 + 18) & 1) != 0)
result = *(_QWORD*)(v3 + 0x20) + 16i64;
else
result = 0i64;
return reinterpret_cast<CStudioHWDataRef*>(result);
}
See description Fixed rare crash in CMDLCache when buffer itteration returns invalid pointer (not null!) for handle. Made all variables and functions used by BuildPropStaticFrustumCullMap searched dynamically (these should also work on s0, s1 and s2, they are compared but at the moment untested). TODO: Fix deref crash properly (see bsplib.cpp l291).
2022-05-01 05:38:51 +02:00
//-----------------------------------------------------------------------------
// Purpose: gets the studio material glue from cache pool by handle
// Input : *this -
// handle -
// Output : a pointer to the CMaterialGlue object
//-----------------------------------------------------------------------------
void* CMDLCache::GetStudioMaterialGlue(CMDLCache* cache, MDLHandle_t handle)
{
__int64 v2; // rbx
__int64 v3; // rbx
v2 = handle;
EnterCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
v3 = *(_QWORD*)(m_MDLDict.Deref().GetPtr() + 24 * v2 + 16);
LeaveCriticalSection(reinterpret_cast<LPCRITICAL_SECTION>(&*m_MDLMutex));
return (char*)v3 + 40;
}
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
void MDLCache_Attach()
{
DetourAttach((LPVOID*)&v_CMDLCache__FindMDL, &CMDLCache::FindMDL);
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
DetourAttach((LPVOID*)&v_CMDLCache__FindCachedMDL, &CMDLCache::FindCachedMDL);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
DetourAttach((LPVOID*)&v_CMDLCache__FindUncachedMDL, &CMDLCache::FindUncachedMDL);
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
DetourAttach((LPVOID*)&v_CMDLCache__GetStudioHardwareRef, &CMDLCache::GetStudioHardwareRef);
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
DetourAttach((LPVOID*)&v_CMDLCache__GetStudioHDR, &CMDLCache::GetStudioHDR);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
void MDLCache_Detach()
{
DetourDetach((LPVOID*)&v_CMDLCache__FindMDL, &CMDLCache::FindMDL);
Add mutex for CMDLCache methods Als fall back to old GatherProps when at least one model is replaced with mdl/error.rmdl as the new GatherProps function does not support this yet.
2022-04-30 03:30:16 +02:00
DetourDetach((LPVOID*)&v_CMDLCache__FindCachedMDL, &CMDLCache::FindCachedMDL);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
DetourDetach((LPVOID*)&v_CMDLCache__FindUncachedMDL, &CMDLCache::FindUncachedMDL);
Early working implementation of the 'mdl/error.rmdl' fallback Only works for prop_static (for the time being) Only works for the S3 game_dll. In order to get this to work: you must set 'old_gather_props' to nun-null and (for certain maps) perform the patch to prevent the frustum culling code to deref a bad pointer (see command 'opcodes test').
2022-04-29 18:25:54 +02:00
DetourDetach((LPVOID*)&v_CMDLCache__GetStudioHardwareRef, &CMDLCache::GetStudioHardwareRef);
Made missing model error handling more robust
2022-04-29 21:05:26 +02:00
DetourDetach((LPVOID*)&v_CMDLCache__GetStudioHDR, &CMDLCache::GetStudioHDR);
Start or 'mdl/error.rmdl' fallback implementation (see description) In 'datacache/mdlcache.cpp' the function 'CMDLCache::FindMDL' attempts to find 'mdl/error.rmdl' and assigns the studiohdr and handle to the members of CMDLFallback. In 'CMDLCache::FindUncachedMDL' we check if a model exists, if a model does not exist, we replace the studiohdr with the one of error.rmdl we stored in the CMDLFallback structure. This does actually work (on the dedicated server it doesn't crash at all!), but on the client it crashes when trying to gather props (right before rendering), setting the ConVar* 'old_gather_props' does interesting things (check IDA around this ConVar*). setting this to 1 causes it to crash in another CMDLCache method when trying to deref something in the global cache pool. This method is easy to rebuild. I will do this soon and attempt to return error.rmdl parts from here as well if results are null (this might actually just work). Leaving 'old_gather_props' to 0 causes it to crash in the middle of the function which is supposed to gather the props the 'new' way. The gather props functions are kind of daunting NOTE: Currently only confirmed to work somewhat properly on the dedicated server for prop_static. prop_dynamic is unconfirmed. And (almost?) works on the client.
2022-04-29 05:30:06 +02:00
}
Reference in New Issue Copy Permalink