`VMA Tracking corruption` is a vengeful spirit I have been getting hit by this hard recently. It seems to occur fairly frequently with Steam when it is hammering a big game download. In my case I'm hitting it with `Spider-Man 2` quite consistently. Backtrace: ``` #0 FEXCore::Assert::ForcedAssert () at /mnt/Work/Work/work/FEXNew/FEXCore/Source/Utils/ForcedAssert.cpp:9 9 asm volatile("hlt #1"); [Current thread is 1 (Thread 0xc1438cc0 (LWP 72746))] (gdb) bt #0 FEXCore::Assert::ForcedAssert () at /mnt/Work/Work/work/FEXNew/FEXCore/Source/Utils/ForcedAssert.cpp:9 #1 0x0000aaaae53d8be0 in LogMan::Throw::MFmt (fmt=, args=...) at /mnt/Work/Work/work/FEXNew/FEXCore/Source/Utils/LogManager.cpp:29 #2 0x0000aaaae52dd964 in LogMan::Throw::AFmt<>(bool, char const*) (Value=, fmt=) at /mnt/Work/Work/work/FEXNew/FEXCore/include/FEXCore/Utils/LogManager.h:53 #3 FEX::HLE::SyscallHandler::TrackMremap (this=0xffff94622000, Thread=0x48e88b000, OldAddress=3100446720, OldSize=, NewSize=, flags=1, NewAddress=2937729024) at /mnt/Work/Work/work/FEXNew/Source/Tools/LinuxEmulation/LinuxSyscalls/SyscallsSMCTracking.cpp:283 #4 0x0000aaaae52ea724 in FEX::HLE::x32::RegisterMemory(FEX::HLE::SyscallHandler*)::$_4::operator()(FEXCore::Core::CpuStateFrame*, void*, unsigned long, unsigned long, int, void*) const (Frame=0x48e88b0a0, old_address=0xb8cd1000, old_size=806912, new_size=1052672, flags=1, new_address=, this=) at /mnt/Work/Work/work/FEXNew/Source/Tools/LinuxEmulation/LinuxSyscalls/x32/Memory.cpp:95 #5 FEX::HLE::x32::RegisterMemory(FEX::HLE::SyscallHandler*)::$_4::__invoke(FEXCore::Core::CpuStateFrame*, void*, unsigned long, unsigned long, int, void*) (Frame=0x48e88b0a0, old_address=0xb8cd1000, old_size=806912, new_size=1052672, flags=1, new_address=) at /mnt/Work/Work/work/FEXNew/Source/Tools/LinuxEmulation/LinuxSyscalls/x32/Memory.cpp:95 #6 0x00000005f53f3ecc in ?? () #7 0x000000048e882000 in ?? () #8 0x0001010000000100 in ?? () ``` Without assertions enabled it of course just results in a crash. Once the game is downloaded, it seems fine enough. The `ThreadedValidation` step doesn't seem to cause problems. This bug has also been around for a few months, so it's not a recent regression.