"Fix" debug console, we were checking the cvar before any cvars were loaded, and the condition it checks in AttachConsole is somehow always false
Remove dead #if 0'd code in math.h On amd64, page_size == 4096 constant, on amd64 w/ win32, allocation_granularity == 65536. These values for x86 windows havent changed over the last 20 years so this is probably safe and gives a modest code size reduction Enable XE_USE_KUSER_SHARED. This sources host time from KUSER_SHARED instead of from QueryPerformanceCounter, which is far faster, but only has a granularity of 100 nanoseconds. In some games seemingly random crashes were happening that were hard to trace because the faulting thread was actually not the one that was misbehaving, another threads stack was underflowing into the faulting thread. Added a bunch of code to synchronize the guest stack and host stack so that if a guest longjmps the host's stack will be adjusted. Changes were also made to allow the guest to call into a piece of an existing x64 function. This synchronization might have a slight performance impact on lower end cpus, to disable it set enable_host_guest_stack_synchronization to false. It is possible it may have introduced regressions, but i dont know of any yet So far, i know the synchronization change fixes the "hub crash" in super sonic and allows the game "london 2012" to go ingame. Removed emit_useless_fpscr_updates, not emitting these updates breaks the raiden game MapGuestAddressToMachineCode now returns nullptr if no address was found, instead of the start of the function add Processor::LookupModule Add Backend::DeinitializeBackendContext Use WriteRegisterRangeFromRing_WithKnownBound<0, 0xFFFF> in WriteRegisterRangeFromRing for inlining (previously regressed on performance of ExecutePacketType0) add notes about flags that trap in XamInputGetCapabilities 0 == 3 in XamInputGetCapabilities Name arg 2 of XamInputSetState PrefetchW in critical section kernel funcs if available & doing cmpxchg Add terminated field to X_KTHREAD, set it on termination Expanded the logic of NtResumeThread/NtSuspendThread to include checking the type of the handle (in release, LookupObject doesnt seem to do anything with the type) and returning X_STATUS_OBJECT_TYPE_MISMATCH if invalid. Do termination check in NtSuspendThread. Add basic host exception messagebox, need to flesh it out more (maybe use the new stack tracking stuff if on guest thrd?) Add rdrand patching hack, mostly affects users with nvidia cards who have many threads on zen Use page_size_shift in more places Once again disable precompilation! Raiden is mostly weird ppc asm which probably breaks the precompilation. The code is still useful for running the compiler over the whole of an xex in debug to test for issues "Fix" debug console, we were checking the cvar before any cvars were loaded, and the condition it checks in AttachConsole is somehow always false Remove dead #if 0'd code in math.h On amd64, page_size == 4096 constant, on amd64 w/ win32, allocation_granularity == 65536. These values for x86 windows havent changed over the last 20 years so this is probably safe and gives a modest code size reduction Enable XE_USE_KUSER_SHARED. This sources host time from KUSER_SHARED instead of from QueryPerformanceCounter, which is far faster, but only has a granularity of 100 nanoseconds. In some games seemingly random crashes were happening that were hard to trace because the faulting thread was actually not the one that was misbehaving, another threads stack was underflowing into the faulting thread. Added a bunch of code to synchronize the guest stack and host stack so that if a guest longjmps the host's stack will be adjusted. Changes were also made to allow the guest to call into a piece of an existing x64 function. This synchronization might have a slight performance impact on lower end cpus, to disable it set enable_host_guest_stack_synchronization to false. It is possible it may have introduced regressions, but i dont know of any yet So far, i know the synchronization change fixes the "hub crash" in super sonic and allows the game "london 2012" to go ingame. Removed emit_useless_fpscr_updates, not emitting these updates breaks the raiden game MapGuestAddressToMachineCode now returns nullptr if no address was found, instead of the start of the function add Processor::LookupModule Add Backend::DeinitializeBackendContext Use WriteRegisterRangeFromRing_WithKnownBound<0, 0xFFFF> in WriteRegisterRangeFromRing for inlining (previously regressed on performance of ExecutePacketType0) add notes about flags that trap in XamInputGetCapabilities 0 == 3 in XamInputGetCapabilities Name arg 2 of XamInputSetState PrefetchW in critical section kernel funcs if available & doing cmpxchg Add terminated field to X_KTHREAD, set it on termination Expanded the logic of NtResumeThread/NtSuspendThread to include checking the type of the handle (in release, LookupObject doesnt seem to do anything with the type) and returning X_STATUS_OBJECT_TYPE_MISMATCH if invalid. Do termination check in NtSuspendThread. Add basic host exception messagebox, need to flesh it out more (maybe use the new stack tracking stuff if on guest thrd?) Add rdrand patching hack, mostly affects users with nvidia cards who have many threads on zen Use page_size_shift in more places Once again disable precompilation! Raiden is mostly weird ppc asm which probably breaks the precompilation. The code is still useful for running the compiler over the whole of an xex in debug to test for issues
This commit is contained in:
chss95cs@gmail.com
@@ -27,18 +27,13 @@
|
||||
#include "xenia/cpu/ppc/ppc_frontend.h"
|
||||
#include "xenia/cpu/ppc/ppc_opcode_info.h"
|
||||
#include "xenia/cpu/processor.h"
|
||||
|
||||
#include "xenia/cpu/xex_module.h"
|
||||
DEFINE_bool(
|
||||
break_on_unimplemented_instructions, true,
|
||||
"Break to the host debugger (or crash if no debugger attached) if an "
|
||||
"unimplemented PowerPC instruction is encountered.",
|
||||
"CPU");
|
||||
|
||||
DEFINE_bool(
|
||||
emit_useless_fpscr_updates, false,
|
||||
"Emit useless fpscr update instructions (pre-10/30/2022 behavior). ",
|
||||
"CPU");
|
||||
|
||||
namespace xe {
|
||||
namespace cpu {
|
||||
namespace ppc {
|
||||
@@ -94,8 +89,9 @@ bool PPCHIRBuilder::Emit(GuestFunction* function, uint32_t flags) {
|
||||
|
||||
function_ = function;
|
||||
start_address_ = function_->address();
|
||||
//chrispy: i've seen this one happen, not sure why but i think from trying to precompile twice
|
||||
//i've also seen ones with a start and end address that are the same...
|
||||
// chrispy: i've seen this one happen, not sure why but i think from trying to
|
||||
// precompile twice i've also seen ones with a start and end address that are
|
||||
// the same...
|
||||
assert_true(function_->address() <= function_->end_address());
|
||||
instr_count_ = (function_->end_address() - function_->address()) / 4 + 1;
|
||||
|
||||
@@ -250,7 +246,8 @@ void PPCHIRBuilder::MaybeBreakOnInstruction(uint32_t address) {
|
||||
}
|
||||
|
||||
void PPCHIRBuilder::AnnotateLabel(uint32_t address, Label* label) {
|
||||
//chrispy: label->name is unused, it would be nice to be able to remove the field and this code
|
||||
// chrispy: label->name is unused, it would be nice to be able to remove the
|
||||
// field and this code
|
||||
char name_buffer[13];
|
||||
auto format_result = fmt::format_to_n(name_buffer, 12, "loc_{:08X}", address);
|
||||
name_buffer[format_result.size] = '\0';
|
||||
@@ -457,37 +454,38 @@ void PPCHIRBuilder::UpdateFPSCR(Value* result, bool update_cr1) {
|
||||
// TODO(benvanik): detect overflow and nan cases.
|
||||
// fx and vx are the most important.
|
||||
/*
|
||||
chrispy: stubbed this out because right now all it does is waste
|
||||
memory and CPU time
|
||||
chrispy: i stubbed this out at one point because all it does is waste
|
||||
memory and CPU time, however, this introduced issues with raiden
|
||||
(substitute w/ titleid later) which probably means they stash stuff in the
|
||||
fpscr?
|
||||
|
||||
*/
|
||||
if (cvars::emit_useless_fpscr_updates) {
|
||||
Value* fx = LoadConstantInt8(0);
|
||||
Value* fex = LoadConstantInt8(0);
|
||||
Value* vx = LoadConstantInt8(0);
|
||||
Value* ox = LoadConstantInt8(0);
|
||||
|
||||
if (update_cr1) {
|
||||
// Store into the CR1 field.
|
||||
// We do this instead of just calling CopyFPSCRToCR1 so that we don't
|
||||
// have to read back the bits and do shifting work.
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_fx), fx);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_fex), fex);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_vx), vx);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_ox), ox);
|
||||
}
|
||||
Value* fx = LoadConstantInt8(0);
|
||||
Value* fex = LoadConstantInt8(0);
|
||||
Value* vx = LoadConstantInt8(0);
|
||||
Value* ox = LoadConstantInt8(0);
|
||||
|
||||
// Generate our new bits.
|
||||
Value* new_bits = Shl(ZeroExtend(fx, INT32_TYPE), 31);
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(fex, INT32_TYPE), 30));
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(vx, INT32_TYPE), 29));
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(ox, INT32_TYPE), 28));
|
||||
|
||||
// Mix into fpscr while preserving sticky bits (FX and OX).
|
||||
Value* bits = LoadFPSCR();
|
||||
bits = Or(And(bits, LoadConstantUint32(0x9FFFFFFF)), new_bits);
|
||||
StoreFPSCR(bits);
|
||||
if (update_cr1) {
|
||||
// Store into the CR1 field.
|
||||
// We do this instead of just calling CopyFPSCRToCR1 so that we don't
|
||||
// have to read back the bits and do shifting work.
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_fx), fx);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_fex), fex);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_vx), vx);
|
||||
StoreContext(offsetof(PPCContext, cr1.cr1_ox), ox);
|
||||
}
|
||||
|
||||
// Generate our new bits.
|
||||
Value* new_bits = Shl(ZeroExtend(fx, INT32_TYPE), 31);
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(fex, INT32_TYPE), 30));
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(vx, INT32_TYPE), 29));
|
||||
new_bits = Or(new_bits, Shl(ZeroExtend(ox, INT32_TYPE), 28));
|
||||
|
||||
// Mix into fpscr while preserving sticky bits (FX and OX).
|
||||
Value* bits = LoadFPSCR();
|
||||
bits = Or(And(bits, LoadConstantUint32(0x9FFFFFFF)), new_bits);
|
||||
StoreFPSCR(bits);
|
||||
}
|
||||
|
||||
void PPCHIRBuilder::CopyFPSCRToCR1() {
|
||||
@@ -587,7 +585,24 @@ void PPCHIRBuilder::StoreReserved(Value* val) {
|
||||
Value* PPCHIRBuilder::LoadReserved() {
|
||||
return LoadContext(offsetof(PPCContext, reserved_val), INT64_TYPE);
|
||||
}
|
||||
void PPCHIRBuilder::SetReturnAddress(Value* value) {
|
||||
/*
|
||||
Record the address as being a possible target of a return. This is
|
||||
needed for longjmp emulation. See x64_emitter.cc's ResolveFunction
|
||||
*/
|
||||
Module* mod = this->function_->module();
|
||||
if (value && value->IsConstant()) {
|
||||
if (mod) {
|
||||
XexModule* xexmod = dynamic_cast<XexModule*>(mod);
|
||||
if (xexmod) {
|
||||
auto flags = xexmod->GetInstructionAddressFlags(value->AsUint32());
|
||||
flags->is_return_site = true;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
HIRBuilder::SetReturnAddress(value);
|
||||
}
|
||||
} // namespace ppc
|
||||
} // namespace cpu
|
||||
} // namespace xe
|
||||
|
||||
@@ -80,7 +80,8 @@ class PPCHIRBuilder : public hir::HIRBuilder {
|
||||
|
||||
void StoreReserved(Value* val);
|
||||
Value* LoadReserved();
|
||||
|
||||
//calls original impl in hirbuilder, but also records the is_return_site bit into flags in the guestmodule
|
||||
void SetReturnAddress(Value* value);
|
||||
private:
|
||||
void MaybeBreakOnInstruction(uint32_t address);
|
||||
void AnnotateLabel(uint32_t address, Label* label);
|
||||
|
||||
Reference in New Issue
Block a user