diff options
Diffstat (limited to '')
-rw-r--r-- | src/core/arm/arm_interface.h | 53 | ||||
-rw-r--r-- | src/core/arm/dynarmic/arm_dynarmic.cpp | 35 | ||||
-rw-r--r-- | src/core/arm/dynarmic/arm_dynarmic.h | 10 | ||||
-rw-r--r-- | src/core/arm/unicorn/arm_unicorn.cpp | 27 | ||||
-rw-r--r-- | src/core/arm/unicorn/arm_unicorn.h | 10 | ||||
-rw-r--r-- | src/core/gdbstub/gdbstub.cpp | 50 | ||||
-rw-r--r-- | src/core/hle/kernel/mutex.cpp | 4 | ||||
-rw-r--r-- | src/core/hle/kernel/svc.cpp | 8 | ||||
-rw-r--r-- | src/core/hle/kernel/thread.cpp | 4 | ||||
-rw-r--r-- | src/core/hle/service/acc/profile_manager.cpp | 5 | ||||
-rw-r--r-- | src/video_core/engines/shader_bytecode.h | 195 | ||||
-rw-r--r-- | src/video_core/renderer_opengl/gl_shader_decompiler.cpp | 162 | ||||
-rw-r--r-- | src/yuzu_cmd/config.cpp | 4 | ||||
-rw-r--r-- | src/yuzu_cmd/default_ini.h | 2 |
14 files changed, 435 insertions, 134 deletions
diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h index 0b2af2a9b..867e34932 100644 --- a/src/core/arm/arm_interface.h +++ b/src/core/arm/arm_interface.h @@ -10,7 +10,7 @@ namespace Core { -/// Generic ARM11 CPU interface +/// Generic ARMv8 CPU interface class ARM_Interface : NonCopyable { public: virtual ~ARM_Interface() {} @@ -19,9 +19,9 @@ public: std::array<u64, 31> cpu_registers; u64 sp; u64 pc; - u64 cpsr; - std::array<u128, 32> fpu_registers; - u64 fpscr; + u64 pstate; + std::array<u128, 32> vector_registers; + u64 fpcr; }; /// Runs the CPU until an event happens @@ -69,42 +69,50 @@ public: */ virtual void SetReg(int index, u64 value) = 0; - virtual u128 GetExtReg(int index) const = 0; - - virtual void SetExtReg(int index, u128 value) = 0; - /** - * Gets the value of a VFP register - * @param index Register index (0-31) - * @return Returns the value in the register + * Gets the value of a specified vector register. + * + * @param index The index of the vector register. + * @return the value within the vector register. */ - virtual u32 GetVFPReg(int index) const = 0; + virtual u128 GetVectorReg(int index) const = 0; /** - * Sets a VFP register to the given value - * @param index Register index (0-31) - * @param value Value to set register to + * Sets a given value into a vector register. + * + * @param index The index of the vector register. + * @param value The new value to place in the register. */ - virtual void SetVFPReg(int index, u32 value) = 0; + virtual void SetVectorReg(int index, u128 value) = 0; /** - * Get the current CPSR register - * @return Returns the value of the CPSR register + * Get the current PSTATE register + * @return Returns the value of the PSTATE register */ - virtual u32 GetCPSR() const = 0; + virtual u32 GetPSTATE() const = 0; /** - * Set the current CPSR register - * @param cpsr Value to set CPSR to + * Set the current PSTATE register + * @param pstate Value to set PSTATE to */ - virtual void SetCPSR(u32 cpsr) = 0; + virtual void SetPSTATE(u32 pstate) = 0; virtual VAddr GetTlsAddress() const = 0; virtual void SetTlsAddress(VAddr address) = 0; + /** + * Gets the value within the TPIDR_EL0 (read/write software thread ID) register. + * + * @return the value within the register. + */ virtual u64 GetTPIDR_EL0() const = 0; + /** + * Sets a new value within the TPIDR_EL0 (read/write software thread ID) register. + * + * @param value The new value to place in the register. + */ virtual void SetTPIDR_EL0(u64 value) = 0; /** @@ -119,6 +127,7 @@ public: */ virtual void LoadContext(const ThreadContext& ctx) = 0; + /// Clears the exclusive monitor's state. virtual void ClearExclusiveState() = 0; /// Prepare core for thread reschedule (if needed to correctly handle state) diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp index 0c175d872..3f072c51f 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic.cpp @@ -194,29 +194,20 @@ void ARM_Dynarmic::SetReg(int index, u64 value) { jit->SetRegister(index, value); } -u128 ARM_Dynarmic::GetExtReg(int index) const { +u128 ARM_Dynarmic::GetVectorReg(int index) const { return jit->GetVector(index); } -void ARM_Dynarmic::SetExtReg(int index, u128 value) { +void ARM_Dynarmic::SetVectorReg(int index, u128 value) { jit->SetVector(index, value); } -u32 ARM_Dynarmic::GetVFPReg(int /*index*/) const { - UNIMPLEMENTED(); - return {}; -} - -void ARM_Dynarmic::SetVFPReg(int /*index*/, u32 /*value*/) { - UNIMPLEMENTED(); -} - -u32 ARM_Dynarmic::GetCPSR() const { +u32 ARM_Dynarmic::GetPSTATE() const { return jit->GetPstate(); } -void ARM_Dynarmic::SetCPSR(u32 cpsr) { - jit->SetPstate(cpsr); +void ARM_Dynarmic::SetPSTATE(u32 pstate) { + jit->SetPstate(pstate); } u64 ARM_Dynarmic::GetTlsAddress() const { @@ -239,18 +230,18 @@ void ARM_Dynarmic::SaveContext(ThreadContext& ctx) { ctx.cpu_registers = jit->GetRegisters(); ctx.sp = jit->GetSP(); ctx.pc = jit->GetPC(); - ctx.cpsr = jit->GetPstate(); - ctx.fpu_registers = jit->GetVectors(); - ctx.fpscr = jit->GetFpcr(); + ctx.pstate = jit->GetPstate(); + ctx.vector_registers = jit->GetVectors(); + ctx.fpcr = jit->GetFpcr(); } void ARM_Dynarmic::LoadContext(const ThreadContext& ctx) { jit->SetRegisters(ctx.cpu_registers); jit->SetSP(ctx.sp); jit->SetPC(ctx.pc); - jit->SetPstate(static_cast<u32>(ctx.cpsr)); - jit->SetVectors(ctx.fpu_registers); - jit->SetFpcr(static_cast<u32>(ctx.fpscr)); + jit->SetPstate(static_cast<u32>(ctx.pstate)); + jit->SetVectors(ctx.vector_registers); + jit->SetFpcr(static_cast<u32>(ctx.fpcr)); } void ARM_Dynarmic::PrepareReschedule() { @@ -304,8 +295,8 @@ bool DynarmicExclusiveMonitor::ExclusiveWrite64(std::size_t core_index, VAddr va bool DynarmicExclusiveMonitor::ExclusiveWrite128(std::size_t core_index, VAddr vaddr, u128 value) { return monitor.DoExclusiveOperation(core_index, vaddr, 16, [&] { - Memory::Write64(vaddr, value[0]); - Memory::Write64(vaddr, value[1]); + Memory::Write64(vaddr + 0, value[0]); + Memory::Write64(vaddr + 8, value[1]); }); } diff --git a/src/core/arm/dynarmic/arm_dynarmic.h b/src/core/arm/dynarmic/arm_dynarmic.h index 56c60c853..e61382d3d 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.h +++ b/src/core/arm/dynarmic/arm_dynarmic.h @@ -29,14 +29,12 @@ public: u64 GetPC() const override; u64 GetReg(int index) const override; void SetReg(int index, u64 value) override; - u128 GetExtReg(int index) const override; - void SetExtReg(int index, u128 value) override; - u32 GetVFPReg(int index) const override; - void SetVFPReg(int index, u32 value) override; - u32 GetCPSR() const override; + u128 GetVectorReg(int index) const override; + void SetVectorReg(int index, u128 value) override; + u32 GetPSTATE() const override; + void SetPSTATE(u32 pstate) override; void Run() override; void Step() override; - void SetCPSR(u32 cpsr) override; VAddr GetTlsAddress() const override; void SetTlsAddress(VAddr address) override; void SetTPIDR_EL0(u64 value) override; diff --git a/src/core/arm/unicorn/arm_unicorn.cpp b/src/core/arm/unicorn/arm_unicorn.cpp index 4e02b7cd4..e218a0b15 100644 --- a/src/core/arm/unicorn/arm_unicorn.cpp +++ b/src/core/arm/unicorn/arm_unicorn.cpp @@ -131,33 +131,24 @@ void ARM_Unicorn::SetReg(int regn, u64 val) { CHECKED(uc_reg_write(uc, treg, &val)); } -u128 ARM_Unicorn::GetExtReg(int /*index*/) const { +u128 ARM_Unicorn::GetVectorReg(int /*index*/) const { UNIMPLEMENTED(); static constexpr u128 res{}; return res; } -void ARM_Unicorn::SetExtReg(int /*index*/, u128 /*value*/) { +void ARM_Unicorn::SetVectorReg(int /*index*/, u128 /*value*/) { UNIMPLEMENTED(); } -u32 ARM_Unicorn::GetVFPReg(int /*index*/) const { - UNIMPLEMENTED(); - return {}; -} - -void ARM_Unicorn::SetVFPReg(int /*index*/, u32 /*value*/) { - UNIMPLEMENTED(); -} - -u32 ARM_Unicorn::GetCPSR() const { +u32 ARM_Unicorn::GetPSTATE() const { u64 nzcv{}; CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &nzcv)); return static_cast<u32>(nzcv); } -void ARM_Unicorn::SetCPSR(u32 cpsr) { - u64 nzcv = cpsr; +void ARM_Unicorn::SetPSTATE(u32 pstate) { + u64 nzcv = pstate; CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &nzcv)); } @@ -219,7 +210,7 @@ void ARM_Unicorn::SaveContext(ThreadContext& ctx) { CHECKED(uc_reg_read(uc, UC_ARM64_REG_SP, &ctx.sp)); CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.cpsr)); + CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); for (auto i = 0; i < 29; ++i) { uregs[i] = UC_ARM64_REG_X0 + i; @@ -234,7 +225,7 @@ void ARM_Unicorn::SaveContext(ThreadContext& ctx) { for (int i = 0; i < 32; ++i) { uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = &ctx.fpu_registers[i]; + tregs[i] = &ctx.vector_registers[i]; } CHECKED(uc_reg_read_batch(uc, uregs, tregs, 32)); @@ -246,7 +237,7 @@ void ARM_Unicorn::LoadContext(const ThreadContext& ctx) { CHECKED(uc_reg_write(uc, UC_ARM64_REG_SP, &ctx.sp)); CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.cpsr)); + CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); for (int i = 0; i < 29; ++i) { uregs[i] = UC_ARM64_REG_X0 + i; @@ -261,7 +252,7 @@ void ARM_Unicorn::LoadContext(const ThreadContext& ctx) { for (auto i = 0; i < 32; ++i) { uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = (void*)&ctx.fpu_registers[i]; + tregs[i] = (void*)&ctx.vector_registers[i]; } CHECKED(uc_reg_write_batch(uc, uregs, tregs, 32)); diff --git a/src/core/arm/unicorn/arm_unicorn.h b/src/core/arm/unicorn/arm_unicorn.h index d6f7cf4ab..75761950b 100644 --- a/src/core/arm/unicorn/arm_unicorn.h +++ b/src/core/arm/unicorn/arm_unicorn.h @@ -22,12 +22,10 @@ public: u64 GetPC() const override; u64 GetReg(int index) const override; void SetReg(int index, u64 value) override; - u128 GetExtReg(int index) const override; - void SetExtReg(int index, u128 value) override; - u32 GetVFPReg(int index) const override; - void SetVFPReg(int index, u32 value) override; - u32 GetCPSR() const override; - void SetCPSR(u32 cpsr) override; + u128 GetVectorReg(int index) const override; + void SetVectorReg(int index, u128 value) override; + u32 GetPSTATE() const override; + void SetPSTATE(u32 pstate) override; VAddr GetTlsAddress() const override; void SetTlsAddress(VAddr address) override; void SetTPIDR_EL0(u64 value) override; diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp index cfaf20a88..1b04f68bf 100644 --- a/src/core/gdbstub/gdbstub.cpp +++ b/src/core/gdbstub/gdbstub.cpp @@ -65,9 +65,9 @@ constexpr u32 MSG_WAITALL = 8; constexpr u32 LR_REGISTER = 30; constexpr u32 SP_REGISTER = 31; constexpr u32 PC_REGISTER = 32; -constexpr u32 CPSR_REGISTER = 33; +constexpr u32 PSTATE_REGISTER = 33; constexpr u32 UC_ARM64_REG_Q0 = 34; -constexpr u32 FPSCR_REGISTER = 66; +constexpr u32 FPCR_REGISTER = 66; // TODO/WiP - Used while working on support for FPU constexpr u32 TODO_DUMMY_REG_997 = 997; @@ -116,7 +116,7 @@ constexpr char target_xml[] = <reg name="pc" bitsize="64" type="code_ptr"/> - <flags id="cpsr_flags" size="4"> + <flags id="pstate_flags" size="4"> <field name="SP" start="0" end="0"/> <field name="" start="1" end="1"/> <field name="EL" start="2" end="3"/> @@ -135,7 +135,7 @@ constexpr char target_xml[] = <field name="Z" start="30" end="30"/> <field name="N" start="31" end="31"/> </flags> - <reg name="cpsr" bitsize="32" type="cpsr_flags"/> + <reg name="pstate" bitsize="32" type="pstate_flags"/> </feature> <feature name="org.gnu.gdb.aarch64.fpu"> </feature> @@ -227,10 +227,10 @@ static u64 RegRead(std::size_t id, Kernel::Thread* thread = nullptr) { return thread->context.sp; } else if (id == PC_REGISTER) { return thread->context.pc; - } else if (id == CPSR_REGISTER) { - return thread->context.cpsr; - } else if (id > CPSR_REGISTER && id < FPSCR_REGISTER) { - return thread->context.fpu_registers[id - UC_ARM64_REG_Q0][0]; + } else if (id == PSTATE_REGISTER) { + return thread->context.pstate; + } else if (id > PSTATE_REGISTER && id < FPCR_REGISTER) { + return thread->context.vector_registers[id - UC_ARM64_REG_Q0][0]; } else { return 0; } @@ -247,10 +247,10 @@ static void RegWrite(std::size_t id, u64 val, Kernel::Thread* thread = nullptr) thread->context.sp = val; } else if (id == PC_REGISTER) { thread->context.pc = val; - } else if (id == CPSR_REGISTER) { - thread->context.cpsr = val; - } else if (id > CPSR_REGISTER && id < FPSCR_REGISTER) { - thread->context.fpu_registers[id - (CPSR_REGISTER + 1)][0] = val; + } else if (id == PSTATE_REGISTER) { + thread->context.pstate = val; + } else if (id > PSTATE_REGISTER && id < FPCR_REGISTER) { + thread->context.vector_registers[id - (PSTATE_REGISTER + 1)][0] = val; } } @@ -781,11 +781,11 @@ static void ReadRegister() { LongToGdbHex(reply, RegRead(id, current_thread)); } else if (id == PC_REGISTER) { LongToGdbHex(reply, RegRead(id, current_thread)); - } else if (id == CPSR_REGISTER) { - IntToGdbHex(reply, (u32)RegRead(id, current_thread)); - } else if (id >= UC_ARM64_REG_Q0 && id < FPSCR_REGISTER) { + } else if (id == PSTATE_REGISTER) { + IntToGdbHex(reply, static_cast<u32>(RegRead(id, current_thread))); + } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { LongToGdbHex(reply, RegRead(id, current_thread)); - } else if (id == FPSCR_REGISTER) { + } else if (id == FPCR_REGISTER) { LongToGdbHex(reply, RegRead(TODO_DUMMY_REG_998, current_thread)); } else { LongToGdbHex(reply, RegRead(TODO_DUMMY_REG_997, current_thread)); @@ -811,7 +811,7 @@ static void ReadRegisters() { bufptr += 16; - IntToGdbHex(bufptr, (u32)RegRead(CPSR_REGISTER, current_thread)); + IntToGdbHex(bufptr, static_cast<u32>(RegRead(PSTATE_REGISTER, current_thread))); bufptr += 8; @@ -843,11 +843,11 @@ static void WriteRegister() { RegWrite(id, GdbHexToLong(buffer_ptr), current_thread); } else if (id == PC_REGISTER) { RegWrite(id, GdbHexToLong(buffer_ptr), current_thread); - } else if (id == CPSR_REGISTER) { + } else if (id == PSTATE_REGISTER) { RegWrite(id, GdbHexToInt(buffer_ptr), current_thread); - } else if (id >= UC_ARM64_REG_Q0 && id < FPSCR_REGISTER) { + } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { RegWrite(id, GdbHexToLong(buffer_ptr), current_thread); - } else if (id == FPSCR_REGISTER) { + } else if (id == FPCR_REGISTER) { RegWrite(TODO_DUMMY_REG_998, GdbHexToLong(buffer_ptr), current_thread); } else { RegWrite(TODO_DUMMY_REG_997, GdbHexToLong(buffer_ptr), current_thread); @@ -866,16 +866,16 @@ static void WriteRegisters() { if (command_buffer[0] != 'G') return SendReply("E01"); - for (u32 i = 0, reg = 0; reg <= FPSCR_REGISTER; i++, reg++) { + for (u32 i = 0, reg = 0; reg <= FPCR_REGISTER; i++, reg++) { if (reg <= SP_REGISTER) { RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread); } else if (reg == PC_REGISTER) { RegWrite(PC_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == CPSR_REGISTER) { - RegWrite(CPSR_REGISTER, GdbHexToInt(buffer_ptr + i * 16), current_thread); - } else if (reg >= UC_ARM64_REG_Q0 && reg < FPSCR_REGISTER) { + } else if (reg == PSTATE_REGISTER) { + RegWrite(PSTATE_REGISTER, GdbHexToInt(buffer_ptr + i * 16), current_thread); + } else if (reg >= UC_ARM64_REG_Q0 && reg < FPCR_REGISTER) { RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == FPSCR_REGISTER) { + } else if (reg == FPCR_REGISTER) { RegWrite(TODO_DUMMY_REG_998, GdbHexToLong(buffer_ptr + i * 16), current_thread); } else { UNIMPLEMENTED(); diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp index 36bf0b677..51f4544be 100644 --- a/src/core/hle/kernel/mutex.cpp +++ b/src/core/hle/kernel/mutex.cpp @@ -62,7 +62,7 @@ ResultCode Mutex::TryAcquire(HandleTable& handle_table, VAddr address, Handle ho Handle requesting_thread_handle) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { - return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidAddress); + return ERR_INVALID_ADDRESS; } SharedPtr<Thread> holding_thread = handle_table.Get<Thread>(holding_thread_handle); @@ -100,7 +100,7 @@ ResultCode Mutex::TryAcquire(HandleTable& handle_table, VAddr address, Handle ho ResultCode Mutex::Release(VAddr address) { // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { - return ResultCode(ErrorModule::Kernel, ErrCodes::InvalidAddress); + return ERR_INVALID_ADDRESS; } auto [thread, num_waiters] = GetHighestPriorityMutexWaitingThread(GetCurrentThread(), address); diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index c5c1697ee..371fc439e 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -280,6 +280,10 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr, "requesting_current_thread_handle=0x{:08X}", holding_thread_handle, mutex_addr, requesting_thread_handle); + if (Memory::IsKernelVirtualAddress(mutex_addr)) { + return ERR_INVALID_ADDRESS_STATE; + } + auto& handle_table = Core::System::GetInstance().Kernel().HandleTable(); return Mutex::TryAcquire(handle_table, mutex_addr, holding_thread_handle, requesting_thread_handle); @@ -289,6 +293,10 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr, static ResultCode ArbitrateUnlock(VAddr mutex_addr) { LOG_TRACE(Kernel_SVC, "called mutex_addr=0x{:X}", mutex_addr); + if (Memory::IsKernelVirtualAddress(mutex_addr)) { + return ERR_INVALID_ADDRESS_STATE; + } + return Mutex::Release(mutex_addr); } diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index 89cd5f401..d4183d6e3 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -217,8 +217,8 @@ static void ResetThreadContext(Core::ARM_Interface::ThreadContext& context, VAdd context.cpu_registers[0] = arg; context.pc = entry_point; context.sp = stack_top; - context.cpsr = 0; - context.fpscr = 0; + context.pstate = 0; + context.fpcr = 0; } ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name, VAddr entry_point, diff --git a/src/core/hle/service/acc/profile_manager.cpp b/src/core/hle/service/acc/profile_manager.cpp index 0071ca613..bcb3475db 100644 --- a/src/core/hle/service/acc/profile_manager.cpp +++ b/src/core/hle/service/acc/profile_manager.cpp @@ -25,7 +25,7 @@ const UUID& UUID::Generate() { ProfileManager::ProfileManager() { // TODO(ogniK): Create the default user we have for now until loading/saving users is added auto user_uuid = UUID{1, 0}; - CreateNewUser(user_uuid, Settings::values.username); + ASSERT(CreateNewUser(user_uuid, Settings::values.username).IsSuccess()); OpenUser(user_uuid); } @@ -91,7 +91,8 @@ ResultCode ProfileManager::CreateNewUser(UUID uuid, const ProfileUsername& usern /// specifically by allowing an std::string for the username. This is required specifically since /// we're loading a string straight from the config ResultCode ProfileManager::CreateNewUser(UUID uuid, const std::string& username) { - ProfileUsername username_output; + ProfileUsername username_output{}; + if (username.size() > username_output.size()) { std::copy_n(username.begin(), username_output.size(), username_output.begin()); } else { diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index 6e555ea03..7e1de0fa1 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -240,6 +240,41 @@ enum class FlowCondition : u64 { Fcsm_Tr = 0x1C, // TODO(bunnei): What is this used for? }; +enum class ControlCode : u64 { + F = 0, + LT = 1, + EQ = 2, + LE = 3, + GT = 4, + NE = 5, + GE = 6, + Num = 7, + Nan = 8, + LTU = 9, + EQU = 10, + LEU = 11, + GTU = 12, + NEU = 13, + GEU = 14, + // + OFF = 16, + LO = 17, + SFF = 18, + LS = 19, + HI = 20, + SFT = 21, + HS = 22, + OFT = 23, + CSM_TA = 24, + CSM_TR = 25, + CSM_MX = 26, + FCSM_TA = 27, + FCSM_TR = 28, + FCSM_MX = 29, + RLE = 30, + RGT = 31, +}; + enum class PredicateResultMode : u64 { None = 0x0, NotZero = 0x3, @@ -271,6 +306,15 @@ enum class TextureProcessMode : u64 { LLA = 7 // Load LOD. The A is unknown, does not appear to differ with LL }; +enum class TextureMiscMode : u64 { + DC, + AOFFI, // Uses Offset + NDV, + NODEP, + MZ, + PTP, +}; + enum class IpaInterpMode : u64 { Linear = 0, Perspective = 1, Flat = 2, Sc = 3 }; enum class IpaSampleMode : u64 { Default = 0, Centroid = 1, Offset = 2 }; @@ -546,6 +590,15 @@ union Instruction { } pset; union { + BitField<0, 3, u64> pred0; + BitField<3, 3, u64> pred3; + BitField<8, 5, ControlCode> cc; // flag in cc + BitField<39, 3, u64> pred39; + BitField<42, 1, u64> neg_pred39; + BitField<45, 4, PredOperation> op; // op with pred39 + } csetp; + + union { BitField<39, 3, u64> pred39; BitField<42, 1, u64> neg_pred; BitField<43, 1, u64> neg_a; @@ -590,42 +643,127 @@ union Instruction { BitField<28, 1, u64> array; BitField<29, 2, TextureType> texture_type; BitField<31, 4, u64> component_mask; + BitField<49, 1, u64> nodep_flag; + BitField<50, 1, u64> dc_flag; + BitField<54, 1, u64> aoffi_flag; BitField<55, 3, TextureProcessMode> process_mode; bool IsComponentEnabled(std::size_t component) const { return ((1ull << component) & component_mask) != 0; } + + TextureProcessMode GetTextureProcessMode() const { + return process_mode; + } + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::DC: + return dc_flag != 0; + case TextureMiscMode::NODEP: + return nodep_flag != 0; + case TextureMiscMode::AOFFI: + return aoffi_flag != 0; + default: + break; + } + return false; + } } tex; union { BitField<22, 6, TextureQueryType> query_type; BitField<31, 4, u64> component_mask; + BitField<49, 1, u64> nodep_flag; + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::NODEP: + return nodep_flag != 0; + default: + break; + } + return false; + } } txq; union { BitField<28, 1, u64> array; BitField<29, 2, TextureType> texture_type; BitField<31, 4, u64> component_mask; + BitField<35, 1, u64> ndv_flag; + BitField<49, 1, u64> nodep_flag; bool IsComponentEnabled(std::size_t component) const { return ((1ull << component) & component_mask) != 0; } + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::NDV: + return (ndv_flag != 0); + case TextureMiscMode::NODEP: + return (nodep_flag != 0); + default: + break; + } + return false; + } } tmml; union { BitField<28, 1, u64> array; BitField<29, 2, TextureType> texture_type; + BitField<35, 1, u64> ndv_flag; + BitField<49, 1, u64> nodep_flag; + BitField<50, 1, u64> dc_flag; + BitField<54, 2, u64> info; BitField<56, 2, u64> component; + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::NDV: + return ndv_flag != 0; + case TextureMiscMode::NODEP: + return nodep_flag != 0; + case TextureMiscMode::DC: + return dc_flag != 0; + case TextureMiscMode::AOFFI: + return info == 1; + case TextureMiscMode::PTP: + return info == 2; + default: + break; + } + return false; + } } tld4; union { + BitField<49, 1, u64> nodep_flag; + BitField<50, 1, u64> dc_flag; + BitField<51, 1, u64> aoffi_flag; BitField<52, 2, u64> component; + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::DC: + return dc_flag != 0; + case TextureMiscMode::NODEP: + return nodep_flag != 0; + case TextureMiscMode::AOFFI: + return aoffi_flag != 0; + default: + break; + } + return false; + } } tld4s; union { BitField<0, 8, Register> gpr0; BitField<28, 8, Register> gpr28; - BitField<49, 1, u64> nodep; + BitField<49, 1, u64> nodep_flag; BitField<50, 3, u64> component_mask_selector; BitField<53, 4, u64> texture_info; @@ -645,6 +783,37 @@ union Instruction { UNREACHABLE(); } + TextureProcessMode GetTextureProcessMode() const { + switch (texture_info) { + case 0: + case 2: + case 6: + case 8: + case 9: + case 11: + return TextureProcessMode::LZ; + case 3: + case 5: + case 13: + return TextureProcessMode::LL; + default: + break; + } + return TextureProcessMode::None; + } + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::DC: + return (texture_info >= 4 && texture_info <= 6) || texture_info == 9; + case TextureMiscMode::NODEP: + return nodep_flag != 0; + default: + break; + } + return false; + } + bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info >= 7 && texture_info <= 9; @@ -673,6 +842,7 @@ union Instruction { } texs; union { + BitField<49, 1, u64> nodep_flag; BitField<53, 4, u64> texture_info; TextureType GetTextureType() const { @@ -693,6 +863,26 @@ union Instruction { UNREACHABLE(); } + TextureProcessMode GetTextureProcessMode() const { + if (texture_info == 1 || texture_info == 5 || texture_info == 12) + return TextureProcessMode::LL; + return TextureProcessMode::LZ; + } + + bool UsesMiscMode(TextureMiscMode mode) const { + switch (mode) { + case TextureMiscMode::AOFFI: + return texture_info == 12 || texture_info == 4; + case TextureMiscMode::MZ: + return texture_info == 5; + case TextureMiscMode::NODEP: + return nodep_flag != 0; + default: + break; + } + return false; + } + bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info == 8; @@ -735,6 +925,7 @@ union Instruction { BitField<36, 5, u64> index; } cbuf36; + BitField<47, 1, u64> generates_cc; BitField<61, 1, u64> is_b_imm; BitField<60, 1, u64> is_b_gpr; BitField<59, 1, u64> is_c_gpr; @@ -859,6 +1050,7 @@ public: ISET_IMM, PSETP, PSET, + CSETP, XMAD_IMM, XMAD_CR, XMAD_RC, @@ -1095,6 +1287,7 @@ private: INST("0011011-0101----", Id::ISET_IMM, Type::IntegerSet, "ISET_IMM"), INST("0101000010001---", Id::PSET, Type::PredicateSetRegister, "PSET"), INST("0101000010010---", Id::PSETP, Type::PredicateSetPredicate, "PSETP"), + INST("010100001010----", Id::CSETP, Type::PredicateSetPredicate, "CSETP"), INST("0011011-00------", Id::XMAD_IMM, Type::Xmad, "XMAD_IMM"), INST("0100111---------", Id::XMAD_CR, Type::Xmad, "XMAD_CR"), INST("010100010-------", Id::XMAD_RC, Type::Xmad, "XMAD_RC"), diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index a1638c12e..b3e95187e 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -236,6 +236,14 @@ private: const std::string& suffix; }; +enum class InternalFlag : u64 { + ZeroFlag = 0, + CarryFlag = 1, + OverflowFlag = 2, + NaNFlag = 3, + Amount +}; + /** * Used to manage shader registers that are emulated with GLSL. This class keeps track of the state * of all registers (e.g. whether they are currently being used as Floats or Integers), and @@ -329,13 +337,19 @@ public: void SetRegisterToInteger(const Register& reg, bool is_signed, u64 elem, const std::string& value, u64 dest_num_components, u64 value_num_components, bool is_saturated = false, - u64 dest_elem = 0, Register::Size size = Register::Size::Word) { + u64 dest_elem = 0, Register::Size size = Register::Size::Word, + bool sets_cc = false) { ASSERT_MSG(!is_saturated, "Unimplemented"); const std::string func{is_signed ? "intBitsToFloat" : "uintBitsToFloat"}; SetRegister(reg, elem, func + '(' + ConvertIntegerSize(value, size) + ')', dest_num_components, value_num_components, dest_elem); + + if (sets_cc) { + const std::string zero_condition = "( " + ConvertIntegerSize(value, size) + " == 0 )"; + SetInternalFlag(InternalFlag::ZeroFlag, zero_condition); + } } /** @@ -352,6 +366,26 @@ public: shader.AddLine(dest + " = " + src + ';'); } + std::string GetControlCode(const Tegra::Shader::ControlCode cc) const { + switch (cc) { + case Tegra::Shader::ControlCode::NEU: + return "!(" + GetInternalFlag(InternalFlag::ZeroFlag) + ')'; + default: + LOG_CRITICAL(HW_GPU, "Unimplemented Control Code {}", static_cast<u32>(cc)); + UNREACHABLE(); + return "false"; + } + } + + std::string GetInternalFlag(const InternalFlag ii) const { + const u32 code = static_cast<u32>(ii); + return "internalFlag_" + std::to_string(code) + suffix; + } + + void SetInternalFlag(const InternalFlag ii, const std::string& value) const { + shader.AddLine(GetInternalFlag(ii) + " = " + value + ';'); + } + /** * Writes code that does a output attribute assignment to register operation. Output attributes * are stored as floats, so this may require conversion. @@ -415,6 +449,12 @@ public: } declarations.AddNewLine(); + for (u32 ii = 0; ii < static_cast<u64>(InternalFlag::Amount); ii++) { + const InternalFlag code = static_cast<InternalFlag>(ii); + declarations.AddLine("bool " + GetInternalFlag(code) + " = false;"); + } + declarations.AddNewLine(); + for (const auto element : declr_input_attribute) { // TODO(bunnei): Use proper number of elements for these u32 idx = @@ -938,8 +978,6 @@ private: // TEXS has two destination registers and a swizzle. The first two elements in the swizzle // go into gpr0+0 and gpr0+1, and the rest goes into gpr28+0 and gpr28+1 - ASSERT_MSG(instr.texs.nodep == 0, "TEXS nodep not implemented"); - std::size_t written_components = 0; for (u32 component = 0; component < 4; ++component) { if (!instr.texs.IsComponentEnabled(component)) { @@ -1622,7 +1660,8 @@ private: } regs.SetRegisterToInteger(instr.gpr0, instr.conversion.is_output_signed, 0, op_a, 1, - 1, instr.alu.saturate_d, 0, instr.conversion.dest_size); + 1, instr.alu.saturate_d, 0, instr.conversion.dest_size, + instr.generates_cc.Value() != 0); break; } case OpCode::Id::I2F_R: @@ -1761,8 +1800,8 @@ private: Tegra::Shader::IpaMode input_mode{Tegra::Shader::IpaInterpMode::Perspective, Tegra::Shader::IpaSampleMode::Default}; - u32 next_element = instr.attribute.fmt20.element; - u32 next_index = static_cast<u32>(instr.attribute.fmt20.index.Value()); + u64 next_element = instr.attribute.fmt20.element; + u64 next_index = static_cast<u64>(instr.attribute.fmt20.index.Value()); const auto LoadNextElement = [&](u32 reg_offset) { regs.SetRegisterToInputAttibute(instr.gpr0.Value() + reg_offset, next_element, @@ -1826,8 +1865,8 @@ private: ASSERT_MSG((instr.attribute.fmt20.immediate.Value() % sizeof(u32)) == 0, "Unaligned attribute loads are not supported"); - u32 next_element = instr.attribute.fmt20.element; - u32 next_index = static_cast<u32>(instr.attribute.fmt20.index.Value()); + u64 next_element = instr.attribute.fmt20.element; + u64 next_index = static_cast<u64>(instr.attribute.fmt20.index.Value()); const auto StoreNextElement = [&](u32 reg_offset) { regs.SetOutputAttributeToRegister(static_cast<Attribute::Index>(next_index), @@ -1853,6 +1892,13 @@ private: Tegra::Shader::TextureType texture_type{instr.tex.texture_type}; std::string coord; + ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI), + "AOFFI is not implemented"); + ASSERT_MSG(!instr.tex.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC), + "DC is not implemented"); + switch (texture_type) { case Tegra::Shader::TextureType::Texture1D: { const std::string x = regs.GetRegisterAsFloat(instr.gpr8); @@ -1935,6 +1981,11 @@ private: Tegra::Shader::TextureType texture_type{instr.texs.GetTextureType()}; bool is_array{instr.texs.IsArrayTexture()}; + ASSERT_MSG(!instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.texs.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC), + "DC is not implemented"); + switch (texture_type) { case Tegra::Shader::TextureType::Texture2D: { if (is_array) { @@ -1971,6 +2022,13 @@ private: ASSERT(instr.tlds.IsArrayTexture() == false); std::string coord; + ASSERT_MSG(!instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI), + "AOFFI is not implemented"); + ASSERT_MSG(!instr.tlds.UsesMiscMode(Tegra::Shader::TextureMiscMode::MZ), + "MZ is not implemented"); + switch (instr.tlds.GetTextureType()) { case Tegra::Shader::TextureType::Texture2D: { if (instr.tlds.IsArrayTexture()) { @@ -1999,6 +2057,17 @@ private: ASSERT(instr.tld4.array == 0); std::string coord; + ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI), + "AOFFI is not implemented"); + ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC), + "DC is not implemented"); + ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV), + "NDV is not implemented"); + ASSERT_MSG(!instr.tld4.UsesMiscMode(Tegra::Shader::TextureMiscMode::PTP), + "PTP is not implemented"); + switch (instr.tld4.texture_type) { case Tegra::Shader::TextureType::Texture2D: { const std::string x = regs.GetRegisterAsFloat(instr.gpr8); @@ -2036,6 +2105,13 @@ private: break; } case OpCode::Id::TLD4S: { + ASSERT_MSG(!instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::AOFFI), + "AOFFI is not implemented"); + ASSERT_MSG(!instr.tld4s.UsesMiscMode(Tegra::Shader::TextureMiscMode::DC), + "DC is not implemented"); + const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8); const std::string op_b = regs.GetRegisterAsFloat(instr.gpr20); // TODO(Subv): Figure out how the sampler type is encoded in the TLD4S instruction. @@ -2048,6 +2124,9 @@ private: break; } case OpCode::Id::TXQ: { + ASSERT_MSG(!instr.txq.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + // TODO: the new commits on the texture refactor, change the way samplers work. // Sadly, not all texture instructions specify the type of texture their sampler // uses. This must be fixed at a later instance. @@ -2068,6 +2147,11 @@ private: break; } case OpCode::Id::TMML: { + ASSERT_MSG(!instr.tmml.UsesMiscMode(Tegra::Shader::TextureMiscMode::NODEP), + "NODEP is not implemented"); + ASSERT_MSG(!instr.tmml.UsesMiscMode(Tegra::Shader::TextureMiscMode::NDV), + "NDV is not implemented"); + const std::string op_a = regs.GetRegisterAsFloat(instr.gpr8); const std::string op_b = regs.GetRegisterAsFloat(instr.gpr8.Value() + 1); const bool is_array = instr.tmml.array != 0; @@ -2234,31 +2318,55 @@ private: break; } case OpCode::Type::PredicateSetPredicate: { - const std::string op_a = - GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0); - const std::string op_b = - GetPredicateCondition(instr.psetp.pred29, instr.psetp.neg_pred29 != 0); + switch (opcode->GetId()) { + case OpCode::Id::PSETP: { + const std::string op_a = + GetPredicateCondition(instr.psetp.pred12, instr.psetp.neg_pred12 != 0); + const std::string op_b = + GetPredicateCondition(instr.psetp.pred29, instr.psetp.neg_pred29 != 0); - // We can't use the constant predicate as destination. - ASSERT(instr.psetp.pred3 != static_cast<u64>(Pred::UnusedIndex)); + // We can't use the constant predicate as destination. + ASSERT(instr.psetp.pred3 != static_cast<u64>(Pred::UnusedIndex)); - const std::string second_pred = - GetPredicateCondition(instr.psetp.pred39, instr.psetp.neg_pred39 != 0); + const std::string second_pred = + GetPredicateCondition(instr.psetp.pred39, instr.psetp.neg_pred39 != 0); - const std::string combiner = GetPredicateCombiner(instr.psetp.op); + const std::string combiner = GetPredicateCombiner(instr.psetp.op); - const std::string predicate = - '(' + op_a + ") " + GetPredicateCombiner(instr.psetp.cond) + " (" + op_b + ')'; + const std::string predicate = + '(' + op_a + ") " + GetPredicateCombiner(instr.psetp.cond) + " (" + op_b + ')'; - // Set the primary predicate to the result of Predicate OP SecondPredicate - SetPredicate(instr.psetp.pred3, - '(' + predicate + ") " + combiner + " (" + second_pred + ')'); + // Set the primary predicate to the result of Predicate OP SecondPredicate + SetPredicate(instr.psetp.pred3, + '(' + predicate + ") " + combiner + " (" + second_pred + ')'); - if (instr.psetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) { - // Set the secondary predicate to the result of !Predicate OP SecondPredicate, - // if enabled - SetPredicate(instr.psetp.pred0, - "!(" + predicate + ") " + combiner + " (" + second_pred + ')'); + if (instr.psetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) { + // Set the secondary predicate to the result of !Predicate OP SecondPredicate, + // if enabled + SetPredicate(instr.psetp.pred0, + "!(" + predicate + ") " + combiner + " (" + second_pred + ')'); + } + break; + } + case OpCode::Id::CSETP: { + const std::string pred = + GetPredicateCondition(instr.csetp.pred39, instr.csetp.neg_pred39 != 0); + const std::string combiner = GetPredicateCombiner(instr.csetp.op); + const std::string controlCode = regs.GetControlCode(instr.csetp.cc); + if (instr.csetp.pred3 != static_cast<u64>(Pred::UnusedIndex)) { + SetPredicate(instr.csetp.pred3, + '(' + controlCode + ") " + combiner + " (" + pred + ')'); + } + if (instr.csetp.pred0 != static_cast<u64>(Pred::UnusedIndex)) { + SetPredicate(instr.csetp.pred0, + "!(" + controlCode + ") " + combiner + " (" + pred + ')'); + } + break; + } + default: { + LOG_CRITICAL(HW_GPU, "Unhandled predicate instruction: {}", opcode->GetName()); + UNREACHABLE(); + } } break; } diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp index 991abda2e..7ec1f5110 100644 --- a/src/yuzu_cmd/config.cpp +++ b/src/yuzu_cmd/config.cpp @@ -125,6 +125,10 @@ void Config::ReadValues() { // System Settings::values.use_docked_mode = sdl2_config->GetBoolean("System", "use_docked_mode", false); + Settings::values.username = sdl2_config->Get("System", "username", "yuzu"); + if (Settings::values.username.empty()) { + Settings::values.username = "yuzu"; + } // Miscellaneous Settings::values.log_filter = sdl2_config->Get("Miscellaneous", "log_filter", "*:Trace"); diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h index 002a4ec15..d35c441e9 100644 --- a/src/yuzu_cmd/default_ini.h +++ b/src/yuzu_cmd/default_ini.h @@ -176,7 +176,7 @@ use_docked_mode = # Sets the account username, max length is 32 characters # yuzu (default) -username = +username = yuzu # Sets the systems language index # 0: Japanese, 1: English (default), 2: French, 3: German, 4: Italian, 5: Spanish, 6: Chinese, |