Address review, clean up code

This commit is contained in:
Louvenarde 2021-08-15 14:10:13 +02:00
parent 8e4c102984
commit f8c177085c
2 changed files with 26 additions and 29 deletions

View File

@ -2,42 +2,38 @@
namespace Components
{
/// This component is a workaround for issue https://github.com/XLabsProject/iw4x-client/issues/80
/// In case the link goes down, this is a "game hangs randomly" issue:
///
/// Investigations on the issue pointed out it comes from a situation on Intel processors where
/// WaitForSingleObjectA is ignored by a thread, for some (?) reason.
///
/// This locks up the game randomly, mostly at the end of rounds or when too many things happen at
/// once, due to trying to stop sounds (AIL_Stop_sounds) and playing streams at the same time,
/// rushing for the same resource via AIL_lock_mutex.
///
/// This bug has been reproduced on mp_terminal, mp_overgrown, mp_rust, with and without bots,
/// and so far this has been the only way to circumvent it afaik. This component wraps
/// miles' mutex into another mutex, created below, and for some reason (?) that mutex is
/// respected when miles' is not.
///
/// As soon as a real fix is found, please discard this fix. In the meantime, it should not
/// have side effects too bad - worst case it might cause a slight performance drop during
/// team switch and intermission.
///
// This component is a workaround for issue https://github.com/XLabsProject/iw4x-client/issues/80
// In case the link goes down, this is a "game hangs randomly" issue:
//
// Investigations on the issue pointed out it comes from a situation on Intel processors where
// WaitForSingleObjectA is ignored by a thread, for some (?) reason.
//
// This locks up the game randomly, mostly at the end of rounds or when too many things happen at
// once, due to trying to stop sounds (AIL_Stop_sounds) and playing streams at the same time,
// rushing for the same resource via AIL_lock_mutex.
//
// This bug has been reproduced on mp_terminal, mp_overgrown, mp_rust, with and without bots,
// and so far this has been the only way to circumvent it afaik. This component wraps
// miles' mutex into another mutex, created below, and for some reason (?) that mutex is
// respected when miles' is not.
//
// As soon as a real fix is found, please discard this fix. In the meantime, it should not
// have side effects too bad - worst case it might cause a slight performance drop during
// team switch and intermission.
//
std::mutex SoundMutexFix::snd_mutex;
static void __stdcall LockSoundMutex(int unk)
void __stdcall SoundMutexFix::LockSoundMutex(int unk)
{
std::lock_guard lock(SoundMutexFix::snd_mutex);
DWORD funcPtr = *reinterpret_cast<DWORD*>(0x6D7554); // AIL_close_stream
((void(__stdcall*)(int unk))(funcPtr))(unk);
Utils::Hook::Call<void __stdcall(int)>(funcPtr)(unk);
}
SoundMutexFix::SoundMutexFix()
{
Utils::Hook(0x689EFE, &LockSoundMutex, HOOK_JUMP).install()->quick();
}
SoundMutexFix::~SoundMutexFix()
{
Utils::Hook(0x689EFE, &SoundMutexFix::LockSoundMutex, HOOK_JUMP).install()->quick();
}
}

View File

@ -1,4 +1,5 @@
#pragma once
#include <mutex>
namespace Components
{
@ -6,9 +7,9 @@ namespace Components
{
public:
SoundMutexFix();
~SoundMutexFix();
static void SND_StopStreamChannelHook(int channel);
private:
static std::mutex snd_mutex;
static void LockSoundMutex(int unk);
};
}