diff options
-rw-r--r-- | .travis-deps.sh | 6 | ||||
-rw-r--r-- | CMakeLists.txt | 30 | ||||
-rw-r--r-- | README.md | 2 | ||||
-rw-r--r-- | src/citra/CMakeLists.txt | 14 | ||||
-rw-r--r-- | src/citra_qt/CMakeLists.txt | 4 | ||||
-rw-r--r-- | src/common/bit_field.h | 6 | ||||
-rw-r--r-- | src/core/arm/dyncom/arm_dyncom_dec.h | 2 | ||||
-rw-r--r-- | src/core/arm/dyncom/arm_dyncom_interpreter.cpp | 106 | ||||
-rw-r--r-- | src/core/arm/dyncom/arm_dyncom_run.cpp | 1 | ||||
-rw-r--r-- | src/core/arm/interpreter/armemu.cpp | 262 | ||||
-rw-r--r-- | src/core/arm/interpreter/armsupp.cpp | 13 | ||||
-rw-r--r-- | src/core/arm/skyeye_common/armdefs.h | 2 | ||||
-rw-r--r-- | src/core/arm/skyeye_common/armemu.h | 8 | ||||
-rw-r--r-- | src/core/hle/service/apt_u.cpp | 4 | ||||
-rw-r--r-- | src/core/hle/service/fs/fs_user.cpp | 19 | ||||
-rw-r--r-- | src/video_core/command_processor.cpp | 4 | ||||
-rw-r--r-- | src/video_core/pica.h | 10 |
17 files changed, 369 insertions, 124 deletions
diff --git a/.travis-deps.sh b/.travis-deps.sh index b978e552e..2a0f6b284 100644 --- a/.travis-deps.sh +++ b/.travis-deps.sh @@ -12,7 +12,11 @@ if [ "$TRAVIS_OS_NAME" = linux -o -z "$TRAVIS_OS_NAME" ]; then ( git clone https://github.com/glfw/glfw.git --branch 3.0.4 --depth 1 mkdir glfw/build && cd glfw/build - cmake .. && make -j2 && sudo make install + cmake -DBUILD_SHARED_LIBS=ON \ + -DGLFW_BUILD_EXAMPLES=OFF \ + -DGLFW_BUILD_TESTS=OFF \ + .. + make -j4 && sudo make install ) sudo apt-get install lib32stdc++6 diff --git a/CMakeLists.txt b/CMakeLists.txt index 638b468a6..5bb87d50d 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -6,18 +6,40 @@ project(citra) if (NOT MSVC) set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wno-attributes") + add_definitions(-pthread) else() # Silence deprecation warnings add_definitions(/D_CRT_SECURE_NO_WARNINGS) # set up output paths for executable binaries (.exe-files, and .dll-files on DLL-capable platforms) set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin) + + # Tweak optimization settings + # As far as I can tell, there's no way to override the CMake defaults while leaving user + # changes intact, so we'll just clobber everything and say sorry. + message(STATUS "Cache compiler flags ignored, please edit CMakeFiles.txt to change the flags.") + # /MD - Multi-threaded runtime + # /Ox - Full optimization + # /Oi - Use intrinsic functions + # /Oy- - Don't omit frame pointer + # /GR- - Disable RTTI + # /GS- - No stack buffer overflow checks + # /EHsc - C++-only exception handling semantics + set(optimization_flags "/MD /Ox /Oi /Oy- /DNDEBUG /GR- /GS- /EHsc") + # /Zi - Output debugging information + # /Zo - enahnced debug info for optimized builds + set(CMAKE_C_FLAGS_RELEASE "${optimization_flags} /Zi" CACHE STRING "" FORCE) + set(CMAKE_CXX_FLAGS_RELEASE "${optimization_flags} /Zi" CACHE STRING "" FORCE) + set(CMAKE_C_FLAGS_RELWITHDEBINFO "${optimization_flags} /Zi /Zo" CACHE STRING "" FORCE) + set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${optimization_flags} /Zi /Zo" CACHE STRING "" FORCE) endif() add_definitions(-DSINGLETHREADED) -find_package(PNG) +find_package(PNG QUIET) if (PNG_FOUND) add_definitions(-DHAVE_PNG) -endif () +else() + message(STATUS "libpng not found. Some debugging features have been disabled.") +endif() find_package(Boost) if (Boost_FOUND) @@ -75,10 +97,6 @@ if (ENABLE_GLFW) set(GLFW_LIBRARIES glfw3) else() - if (NOT APPLE) - find_package(X11 REQUIRED) - endif() - find_package(PkgConfig REQUIRED) pkg_search_module(GLFW REQUIRED glfw3) endif() @@ -10,7 +10,7 @@ For development discussion, please join us @ #citra on [freenode](http://webchat ### Development -If you want to contribute please take a took at the [Contributor's Guide](CONTRIBUTING.md), [Roadmap](https://github.com/citra-emu/citra/wiki/Roadmap) and [Developer Information](https://github.com/citra-emu/citra/wiki/Developer-Information) pages. You should as well contact any of the developers in the forum in order to know about the current state of the emulator. +If you want to contribute please take a look at the [Contributor's Guide](CONTRIBUTING.md), [Roadmap](https://github.com/citra-emu/citra/wiki/Roadmap) and [Developer Information](https://github.com/citra-emu/citra/wiki/Developer-Information) pages. You should as well contact any of the developers in the forum in order to know about the current state of the emulator. ### Building diff --git a/src/citra/CMakeLists.txt b/src/citra/CMakeLists.txt index f2add394f..b06259f5e 100644 --- a/src/citra/CMakeLists.txt +++ b/src/citra/CMakeLists.txt @@ -12,22 +12,20 @@ set(HEADERS create_directory_groups(${SRCS} ${HEADERS}) -# NOTE: This is a workaround for CMake bug 0006976 (missing X11_xf86vmode_LIB variable) -if (NOT X11_xf86vmode_LIB) - set(X11_xv86vmode_LIB Xxf86vm) -endif() - add_executable(citra ${SRCS} ${HEADERS}) target_link_libraries(citra core common video_core) target_link_libraries(citra ${OPENGL_gl_LIBRARY} ${GLFW_LIBRARIES} inih) +if (UNIX) + target_link_libraries(citra -pthread) +endif() + if (APPLE) - target_link_libraries(citra iconv pthread ${COREFOUNDATION_LIBRARY}) + target_link_libraries(citra iconv ${COREFOUNDATION_LIBRARY}) elseif (WIN32) target_link_libraries(citra winmm) else() # Unix - target_link_libraries(citra pthread rt) - target_link_libraries(citra ${X11_X11_LIB} ${X11_Xi_LIB} ${X11_Xcursor_LIB} ${X11_Xrandr_LIB} ${X11_xv86vmode_LIB}) + target_link_libraries(citra rt) endif() #install(TARGETS citra RUNTIME DESTINATION ${bindir}) diff --git a/src/citra_qt/CMakeLists.txt b/src/citra_qt/CMakeLists.txt index 90e5c6aa6..54d0a1271 100644 --- a/src/citra_qt/CMakeLists.txt +++ b/src/citra_qt/CMakeLists.txt @@ -60,6 +60,10 @@ add_executable(citra-qt ${SRCS} ${HEADERS} ${UI_HDRS}) target_link_libraries(citra-qt core common video_core qhexedit) target_link_libraries(citra-qt ${OPENGL_gl_LIBRARY} ${CITRA_QT_LIBS}) +if (UNIX) + target_link_libraries(citra-qt -pthread) +endif() + if (APPLE) target_link_libraries(citra-qt iconv ${COREFOUNDATION_LIBRARY}) elseif (WIN32) diff --git a/src/common/bit_field.h b/src/common/bit_field.h index c38a3fb08..8eab054b8 100644 --- a/src/common/bit_field.h +++ b/src/common/bit_field.h @@ -168,6 +168,12 @@ public: } } + // TODO: we may want to change this to explicit operator bool() if it's bug-free in VS2015 + __forceinline bool ToBool() const + { + return Value() != 0; + } + private: // StorageType is T for non-enum types and the underlying type of T if // T is an enumeration. Note that T is wrapped within an enable_if in the diff --git a/src/core/arm/dyncom/arm_dyncom_dec.h b/src/core/arm/dyncom/arm_dyncom_dec.h index 19d94f369..70eb96e93 100644 --- a/src/core/arm/dyncom/arm_dyncom_dec.h +++ b/src/core/arm/dyncom/arm_dyncom_dec.h @@ -56,8 +56,6 @@ #define RN ((instr >> 16) & 0xF) /*xxxx xxxx xxxx xxxx xxxx xxxx xxxx 1111 */ #define RM (instr & 0xF) -#define BIT(n) ((instr >> (n)) & 1) -#define BITS(a,b) ((instr >> (a)) & ((1 << (1+(b)-(a)))-1)) /* CP15 registers */ #define OPCODE_1 BITS(21, 23) diff --git a/src/core/arm/dyncom/arm_dyncom_interpreter.cpp b/src/core/arm/dyncom/arm_dyncom_interpreter.cpp index df698e8f1..460001b1a 100644 --- a/src/core/arm/dyncom/arm_dyncom_interpreter.cpp +++ b/src/core/arm/dyncom/arm_dyncom_interpreter.cpp @@ -2390,15 +2390,41 @@ ARM_INST_PTR INTERPRETER_TRANSLATE(pld)(unsigned int inst, int index) return inst_base; } ARM_INST_PTR INTERPRETER_TRANSLATE(qadd)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QADD"); } -ARM_INST_PTR INTERPRETER_TRANSLATE(qadd16)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QADD16"); } ARM_INST_PTR INTERPRETER_TRANSLATE(qadd8)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QADD8"); } -ARM_INST_PTR INTERPRETER_TRANSLATE(qaddsubx)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QADDSUBX"); } +ARM_INST_PTR INTERPRETER_TRANSLATE(qadd16)(unsigned int inst, int index) +{ + arm_inst* const inst_base = (arm_inst*)AllocBuffer(sizeof(arm_inst) + sizeof(generic_arm_inst)); + generic_arm_inst* const inst_cream = (generic_arm_inst*)inst_base->component; + + inst_base->cond = BITS(inst, 28, 31); + inst_base->idx = index; + inst_base->br = NON_BRANCH; + inst_base->load_r15 = 0; + + inst_cream->Rm = BITS(inst, 0, 3); + inst_cream->Rn = BITS(inst, 16, 19); + inst_cream->Rd = BITS(inst, 12, 15); + inst_cream->op1 = BITS(inst, 20, 21); + inst_cream->op2 = BITS(inst, 5, 7); + + return inst_base; +} +ARM_INST_PTR INTERPRETER_TRANSLATE(qaddsubx)(unsigned int inst, int index) +{ + return INTERPRETER_TRANSLATE(qadd16)(inst, index); +} ARM_INST_PTR INTERPRETER_TRANSLATE(qdadd)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QDADD"); } ARM_INST_PTR INTERPRETER_TRANSLATE(qdsub)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QDSUB"); } ARM_INST_PTR INTERPRETER_TRANSLATE(qsub)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QSUB"); } -ARM_INST_PTR INTERPRETER_TRANSLATE(qsub16)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QSUB16"); } ARM_INST_PTR INTERPRETER_TRANSLATE(qsub8)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QSUB8"); } -ARM_INST_PTR INTERPRETER_TRANSLATE(qsubaddx)(unsigned int inst, int index) { UNIMPLEMENTED_INSTRUCTION("QSUBADDX"); } +ARM_INST_PTR INTERPRETER_TRANSLATE(qsub16)(unsigned int inst, int index) +{ + return INTERPRETER_TRANSLATE(qadd16)(inst, index); +} +ARM_INST_PTR INTERPRETER_TRANSLATE(qsubaddx)(unsigned int inst, int index) +{ + return INTERPRETER_TRANSLATE(qadd16)(inst, index); +} ARM_INST_PTR INTERPRETER_TRANSLATE(rev)(unsigned int inst, int index) { arm_inst *inst_base = (arm_inst *)AllocBuffer(sizeof(arm_inst) + sizeof(rev_inst)); @@ -3798,9 +3824,9 @@ unsigned InterpreterMainLoop(ARMul_State* state) #define INC_ICOUNTER cpu->icounter++; \ if(cpu->Reg[15] > 0xc0000000) \ cpu->kernel_icounter++; - //if (debug_function(core)) \ + /*if (debug_function(core)) \ if (core->check_int_flag) \ - goto END + goto END*/ //LOG_TRACE(Core_ARM11, "icounter is %llx pc is %x\n", cpu->icounter, cpu->Reg[15]) #else #define INC_ICOUNTER ; @@ -4021,18 +4047,18 @@ unsigned InterpreterMainLoop(ARMul_State* state) #define UPDATE_NFLAG(dst) (cpu->NFlag = BIT(dst, 31) ? 1 : 0) #define UPDATE_ZFLAG(dst) (cpu->ZFlag = dst ? 0 : 1) -// #define UPDATE_CFLAG(dst, lop, rop) (cpu->CFlag = ((ISNEG(lop) && ISPOS(rop)) || \ +/* #define UPDATE_CFLAG(dst, lop, rop) (cpu->CFlag = ((ISNEG(lop) && ISPOS(rop)) || \ (ISNEG(lop) && ISPOS(dst)) || \ - (ISPOS(rop) && ISPOS(dst)))) + (ISPOS(rop) && ISPOS(dst)))) */ #define UPDATE_CFLAG(dst, lop, rop) (cpu->CFlag = ((dst < lop) || (dst < rop))) #define UPDATE_CFLAG_CARRY_FROM_ADD(lop, rop, flag) (cpu->CFlag = (((uint64_t) lop + (uint64_t) rop + (uint64_t) flag) > 0xffffffff) ) #define UPDATE_CFLAG_NOT_BORROW_FROM_FLAG(lop, rop, flag) (cpu->CFlag = ((uint64_t) lop >= ((uint64_t) rop + (uint64_t) flag))) #define UPDATE_CFLAG_NOT_BORROW_FROM(lop, rop) (cpu->CFlag = (lop >= rop)) #define UPDATE_CFLAG_WITH_NOT(dst, lop, rop) (cpu->CFlag = !(dst < lop)) #define UPDATE_CFLAG_WITH_SC cpu->CFlag = cpu->shifter_carry_out -// #define UPDATE_CFLAG_WITH_NOT(dst, lop, rop) cpu->CFlag = !((ISNEG(lop) && ISPOS(rop)) || \ +/* #define UPDATE_CFLAG_WITH_NOT(dst, lop, rop) cpu->CFlag = !((ISNEG(lop) && ISPOS(rop)) || \ (ISNEG(lop) && ISPOS(dst)) || \ - (ISPOS(rop) && ISPOS(dst))) + (ISPOS(rop) && ISPOS(dst))) */ #define UPDATE_VFLAG(dst, lop, rop) (cpu->VFlag = (((lop < 0) && (rop < 0) && (dst >= 0)) || \ ((lop >= 0) && (rop) >= 0 && (dst < 0)))) #define UPDATE_VFLAG_WITH_NOT(dst, lop, rop) (cpu->VFlag = !(((lop < 0) && (rop < 0) && (dst >= 0)) || \ @@ -5561,15 +5587,69 @@ unsigned InterpreterMainLoop(ARMul_State* state) GOTO_NEXT_INST; } QADD_INST: - QADD16_INST: QADD8_INST: + + QADD16_INST: QADDSUBX_INST: + QSUB16_INST: + QSUBADDX_INST: + { + INC_ICOUNTER; + if (inst_base->cond == 0xE || CondPassed(cpu, inst_base->cond)) { + generic_arm_inst* const inst_cream = (generic_arm_inst*)inst_base->component; + const s16 rm_lo = (RM & 0xFFFF); + const s16 rm_hi = ((RM >> 16) & 0xFFFF); + const s16 rn_lo = (RN & 0xFFFF); + const s16 rn_hi = ((RN >> 16) & 0xFFFF); + const u8 op2 = inst_cream->op2; + + s32 lo_result = 0; + s32 hi_result = 0; + + // QADD16 + if (op2 == 0x00) { + lo_result = (rn_lo + rm_lo); + hi_result = (rn_hi + rm_hi); + } + // QASX + else if (op2 == 0x01) { + lo_result = (rn_lo - rm_hi); + hi_result = (rn_hi + rm_lo); + } + // QSAX + else if (op2 == 0x02) { + lo_result = (rn_lo + rm_hi); + hi_result = (rn_hi - rm_lo); + } + // QSUB16 + else if (op2 == 0x03) { + lo_result = (rn_lo - rm_lo); + hi_result = (rn_hi - rm_hi); + } + + if (lo_result > 0x7FFF) + lo_result = 0x7FFF; + else if (lo_result < -0x8000) + lo_result = -0x8000; + + if (hi_result > 0x7FFF) + hi_result = 0x7FFF; + else if (hi_result < -0x8000) + hi_result = -0x8000; + + RD = (lo_result & 0xFFFF) | ((hi_result & 0xFFFF) << 16); + } + + cpu->Reg[15] += GET_INST_SIZE(cpu); + INC_PC(sizeof(generic_arm_inst)); + FETCH_INST; + GOTO_NEXT_INST; + } + QDADD_INST: QDSUB_INST: QSUB_INST: - QSUB16_INST: QSUB8_INST: - QSUBADDX_INST: REV_INST: { INC_ICOUNTER; diff --git a/src/core/arm/dyncom/arm_dyncom_run.cpp b/src/core/arm/dyncom/arm_dyncom_run.cpp index a2026cbf3..b66b92cf5 100644 --- a/src/core/arm/dyncom/arm_dyncom_run.cpp +++ b/src/core/arm/dyncom/arm_dyncom_run.cpp @@ -29,7 +29,6 @@ void switch_mode(arm_core_t *core, uint32_t mode) { - uint32_t tmp1, tmp2; if (core->Mode == mode) { //Mode not changed. //printf("mode not changed\n"); diff --git a/src/core/arm/interpreter/armemu.cpp b/src/core/arm/interpreter/armemu.cpp index 610e04f10..d54dbeac5 100644 --- a/src/core/arm/interpreter/armemu.cpp +++ b/src/core/arm/interpreter/armemu.cpp @@ -1670,7 +1670,7 @@ mainswitch: op1 *= op2; //printf("SMLA_INST:BB,op1=0x%x, op2=0x%x. Rn=0x%x\n", op1, op2, Rn); if (AddOverflow(op1, Rn, op1 + Rn)) - SETS; + SETQ; state->Reg[BITS (16, 19)] = op1 + Rn; break; } @@ -1682,7 +1682,7 @@ mainswitch: ARMword result = op1 + op2; if (AddOverflow(op1, op2, result)) { result = POS (result) ? 0x80000000 : 0x7fffffff; - SETS; + SETQ; } state->Reg[BITS (12, 15)] = result; break; @@ -1724,7 +1724,7 @@ mainswitch: TAKEABORT; } else if ((BITS (0, 11) == 0) && (LHSReg == 15)) { /* MRS CPSR */ UNDEF_MRSPC; - DEST = ECC | EINT | EMODE; + DEST = ARMul_GetCPSR(state); } else { UNDEF_Test; } @@ -1795,7 +1795,7 @@ mainswitch: ARMword Rn = state->Reg[BITS(12, 15)]; if (AddOverflow((ARMword)result, Rn, (ARMword)(result + Rn))) - SETS; + SETQ; result += Rn; } state->Reg[BITS (16, 19)] = (ARMword)result; @@ -1811,7 +1811,7 @@ mainswitch: if (SubOverflow (op1, op2, result)) { result = POS (result) ? 0x80000000 : 0x7fffffff; - SETS; + SETQ; } state->Reg[BITS (12, 15)] = result; @@ -1934,13 +1934,13 @@ mainswitch: if (AddOverflow (op2, op2, op2d)) { - SETS; + SETQ; op2d = POS (op2d) ? 0x80000000 : 0x7fffffff; } result = op1 + op2d; if (AddOverflow(op1, op2d, result)) { - SETS; + SETQ; result = POS (result) ? 0x80000000 : 0x7fffffff; } @@ -2053,13 +2053,13 @@ mainswitch: ARMword result; if (AddOverflow(op2, op2, op2d)) { - SETS; + SETQ; op2d = POS (op2d) ? 0x80000000 : 0x7fffffff; } result = op1 - op2d; if (SubOverflow(op1, op2d, result)) { - SETS; + SETQ; result = POS (result) ? 0x80000000 : 0x7fffffff; } @@ -5863,20 +5863,21 @@ L_stm_s_takeabort: state->Reg[rd_idx] = (lo_result & 0xFFFF) | ((hi_result & 0xFFFF) << 16); if (lo_result >= 0) { - state->Cpsr |= (1 << 16); - state->Cpsr |= (1 << 17); + state->GEFlag |= (1 << 16); + state->GEFlag |= (1 << 17); } else { - state->Cpsr &= ~(1 << 16); - state->Cpsr &= ~(1 << 17); + state->GEFlag &= ~(1 << 16); + state->GEFlag &= ~(1 << 17); } if (hi_result >= 0) { - state->Cpsr |= (1 << 18); - state->Cpsr |= (1 << 19); + state->GEFlag |= (1 << 18); + state->GEFlag |= (1 << 19); } else { - state->Cpsr &= ~(1 << 18); - state->Cpsr &= ~(1 << 19); + state->GEFlag &= ~(1 << 18); + state->GEFlag &= ~(1 << 19); } + return 1; } // SADD8/SSUB8 @@ -5901,24 +5902,24 @@ L_stm_s_takeabort: hi_val2 = (u8)(((rn_val >> 24) & 0xFF) + ((rm_val >> 24) & 0xFF)); if (lo_val1 & 0x80) - state->Cpsr |= (1 << 16); + state->GEFlag |= (1 << 16); else - state->Cpsr &= ~(1 << 16); + state->GEFlag &= ~(1 << 16); if (lo_val2 & 0x80) - state->Cpsr |= (1 << 17); + state->GEFlag |= (1 << 17); else - state->Cpsr &= ~(1 << 17); + state->GEFlag &= ~(1 << 17); if (hi_val1 & 0x80) - state->Cpsr |= (1 << 18); + state->GEFlag |= (1 << 18); else - state->Cpsr &= ~(1 << 18); + state->GEFlag &= ~(1 << 18); if (hi_val2 & 0x80) - state->Cpsr |= (1 << 19); + state->GEFlag |= (1 << 19); else - state->Cpsr &= ~(1 << 19); + state->GEFlag &= ~(1 << 19); } // SSUB8 else { @@ -5928,24 +5929,24 @@ L_stm_s_takeabort: hi_val2 = (u8)(((rn_val >> 24) & 0xFF) - ((rm_val >> 24) & 0xFF)); if (!(lo_val1 & 0x80)) - state->Cpsr |= (1 << 16); + state->GEFlag |= (1 << 16); else - state->Cpsr &= ~(1 << 16); + state->GEFlag &= ~(1 << 16); if (!(lo_val2 & 0x80)) - state->Cpsr |= (1 << 17); + state->GEFlag |= (1 << 17); else - state->Cpsr &= ~(1 << 17); + state->GEFlag &= ~(1 << 17); if (!(hi_val1 & 0x80)) - state->Cpsr |= (1 << 18); + state->GEFlag |= (1 << 18); else - state->Cpsr &= ~(1 << 18); + state->GEFlag &= ~(1 << 18); if (!(hi_val2 & 0x80)) - state->Cpsr |= (1 << 19); + state->GEFlag |= (1 << 19); else - state->Cpsr &= ~(1 << 19); + state->GEFlag &= ~(1 << 19); } state->Reg[rd_idx] = (lo_val1 | lo_val2 << 8 | hi_val1 << 16 | hi_val2 << 24); @@ -6024,15 +6025,32 @@ L_stm_s_takeabort: if ((instr & 0x0F0) == 0x070) { // USUB16 h1 = ((u16)from - (u16)to); h2 = ((u16)(from >> 16) - (u16)(to >> 16)); - if (!(h1 & 0xffff0000)) state->Cpsr |= (3 << 16); - if (!(h2 & 0xffff0000)) state->Cpsr |= (3 << 18); + + if (!(h1 & 0xffff0000)) + state->GEFlag |= (3 << 16); + else + state->GEFlag &= ~(3 << 16); + + if (!(h2 & 0xffff0000)) + state->GEFlag |= (3 << 18); + else + state->GEFlag &= ~(3 << 18); } else { // UADD16 h1 = ((u16)from + (u16)to); h2 = ((u16)(from >> 16) + (u16)(to >> 16)); - if (h1 & 0xffff0000) state->Cpsr |= (3 << 16); - if (h2 & 0xffff0000) state->Cpsr |= (3 << 18); + + if (h1 & 0xffff0000) + state->GEFlag |= (3 << 16); + else + state->GEFlag &= ~(3 << 16); + + if (h2 & 0xffff0000) + state->GEFlag |= (3 << 18); + else + state->GEFlag &= ~(3 << 18); } + state->Reg[rd] = (u32)((h1 & 0xffff) | ((h2 & 0xffff) << 16)); return 1; } @@ -6045,10 +6063,26 @@ L_stm_s_takeabort: b2 = ((u8)(from >> 8) - (u8)(to >> 8)); b3 = ((u8)(from >> 16) - (u8)(to >> 16)); b4 = ((u8)(from >> 24) - (u8)(to >> 24)); - if (!(b1 & 0xffffff00)) state->Cpsr |= (1 << 16); - if (!(b2 & 0xffffff00)) state->Cpsr |= (1 << 17); - if (!(b3 & 0xffffff00)) state->Cpsr |= (1 << 18); - if (!(b4 & 0xffffff00)) state->Cpsr |= (1 << 19); + + if (!(b1 & 0xffffff00)) + state->GEFlag |= (1 << 16); + else + state->GEFlag &= ~(1 << 16); + + if (!(b2 & 0xffffff00)) + state->GEFlag |= (1 << 17); + else + state->GEFlag &= ~(1 << 17); + + if (!(b3 & 0xffffff00)) + state->GEFlag |= (1 << 18); + else + state->GEFlag &= ~(1 << 18); + + if (!(b4 & 0xffffff00)) + state->GEFlag |= (1 << 19); + else + state->GEFlag &= ~(1 << 19); } else { // UADD8 b1 = ((u8)from + (u8)to); @@ -6057,25 +6091,24 @@ L_stm_s_takeabort: b4 = ((u8)(from >> 24) + (u8)(to >> 24)); if (b1 & 0xffffff00) - state->Cpsr |= (1 << 16); + state->GEFlag |= (1 << 16); else - state->Cpsr &= ~(1 << 16); + state->GEFlag &= ~(1 << 16); if (b2 & 0xffffff00) - state->Cpsr |= (1 << 17); + state->GEFlag |= (1 << 17); else - state->Cpsr &= ~(1 << 17); + state->GEFlag &= ~(1 << 17); if (b3 & 0xffffff00) - state->Cpsr |= (1 << 18); + state->GEFlag |= (1 << 18); else - state->Cpsr &= ~(1 << 18); - + state->GEFlag &= ~(1 << 18); if (b4 & 0xffffff00) - state->Cpsr |= (1 << 19); + state->GEFlag |= (1 << 19); else - state->Cpsr &= ~(1 << 19); + state->GEFlag &= ~(1 << 19); } state->Reg[rd] = (u32)(b1 | (b2 & 0xff) << 8 | (b3 & 0xff) << 16 | (b4 & 0xff) << 24); @@ -6116,7 +6149,7 @@ L_stm_s_takeabort: u32 rm = (instr >> 0) & 0xF; u32 from = state->Reg[rn]; u32 to = state->Reg[rm]; - u32 cpsr = state->Cpsr; + u32 cpsr = ARMul_GetCPSR(state); if ((instr & 0xFF0) == 0xFB0) { // SEL u32 result; if (cpsr & (1 << 16)) @@ -6172,15 +6205,21 @@ L_stm_s_takeabort: s16 rn_lo = (state->Reg[rn_idx]); s16 rn_hi = (state->Reg[rn_idx] >> 16); - if (rn_lo > max) + if (rn_lo > max) { rn_lo = max; - else if (rn_lo < min) + SETQ; + } else if (rn_lo < min) { rn_lo = min; + SETQ; + } - if (rn_hi > max) + if (rn_hi > max) { rn_hi = max; - else if (rn_hi < min) + SETQ; + } else if (rn_hi < min) { rn_hi = min; + SETQ; + } state->Reg[rd_idx] = (rn_lo & 0xFFFF) | ((rn_hi & 0xFFFF) << 16); return 1; @@ -6313,16 +6352,22 @@ L_stm_s_takeabort: s16 rn_lo = (state->Reg[rn_idx]); s16 rn_hi = (state->Reg[rn_idx] >> 16); - if (max < rn_lo) + if (max < rn_lo) { rn_lo = max; - else if (rn_lo < 0) + SETQ; + } else if (rn_lo < 0) { rn_lo = 0; + SETQ; + } - if (max < rn_hi) + if (max < rn_hi) { rn_hi = max; - else if (rn_hi < 0) + SETQ; + } else if (rn_hi < 0) { rn_hi = 0; - + SETQ; + } + state->Reg[rd_idx] = (rn_lo & 0xFFFF) | ((rn_hi << 16) & 0xFFFF); return 1; } @@ -6427,30 +6472,95 @@ L_stm_s_takeabort: const s16 rn_lo = (rn_val & 0xFFFF); const s16 rn_hi = ((rn_val >> 16) & 0xFFFF); - // SMUAD - if ((instr & 0xf0d0) == 0xf010) { - state->Reg[rd_idx] = (rn_lo * rm_lo) + (rn_hi * rm_hi); + const u32 product1 = (rn_lo * rm_lo); + const u32 product2 = (rn_hi * rm_hi); + + // SMUAD and SMLAD + if (BIT(6) == 0) { + state->Reg[rd_idx] = product1 + product2; + + if (BITS(12, 15) != 15) { + state->Reg[rd_idx] += state->Reg[ra_idx]; + ARMul_AddOverflowQ(state, product1 + product2, state->Reg[ra_idx]); + } + + ARMul_AddOverflowQ(state, product1, product2); } - // SMUSD - else if ((instr & 0xf0d0) == 0xf050) { - state->Reg[rd_idx] = (rn_lo * rm_lo) - (rn_hi * rm_hi); + // SMUSD and SMLSD + else { + state->Reg[rd_idx] = product1 - product2; + + if (BITS(12, 15) != 15) + state->Reg[rd_idx] += state->Reg[ra_idx]; } - // SMLAD - else if ((instr & 0xd0) == 0x10) { - state->Reg[rd_idx] = (rn_lo * rm_lo) + (rn_hi * rm_hi) + (s32)state->Reg[ra_idx]; + + return 1; + } + break; + case 0x74: // SMLALD and SMLSLD + { + const u8 rm_idx = BITS(8, 11); + const u8 rn_idx = BITS(0, 3); + const u8 rdlo_idx = BITS(12, 15); + const u8 rdhi_idx = BITS(16, 19); + const bool do_swap = (BIT(5) == 1); + + const u32 rdlo_val = state->Reg[rdlo_idx]; + const u32 rdhi_val = state->Reg[rdhi_idx]; + const u32 rn_val = state->Reg[rn_idx]; + u32 rm_val = state->Reg[rm_idx]; + + if (do_swap) + rm_val = (((rm_val & 0xFFFF) << 16) | (rm_val >> 16)); + + const s32 product1 = (s16)(rn_val & 0xFFFF) * (s16)(rm_val & 0xFFFF); + const s32 product2 = (s16)((rn_val >> 16) & 0xFFFF) * (s16)((rm_val >> 16) & 0xFFFF); + s64 result; + + // SMLALD + if (BIT(6) == 0) { + result = (product1 + product2) + (s64)(rdlo_val | ((s64)rdhi_val << 32)); } - // SMLSD + // SMLSLD else { - state->Reg[rd_idx] = ((rn_lo * rm_lo) - (rn_hi * rm_hi)) + (s32)state->Reg[ra_idx]; + result = (product1 - product2) + (s64)(rdlo_val | ((s64)rdhi_val << 32)); } + + state->Reg[rdlo_idx] = (result & 0xFFFFFFFF); + state->Reg[rdhi_idx] = ((result >> 32) & 0xFFFFFFFF); return 1; } break; - case 0x74: - printf ("Unhandled v6 insn: smlald/smlsld\n"); - break; - case 0x75: - printf ("Unhandled v6 insn: smmla/smmls/smmul\n"); + case 0x75: // SMMLA, SMMUL, and SMMLS + { + const u8 rm_idx = BITS(8, 11); + const u8 rn_idx = BITS(0, 3); + const u8 ra_idx = BITS(12, 15); + const u8 rd_idx = BITS(16, 19); + const bool do_round = (BIT(5) == 1); + + const u32 rm_val = state->Reg[rm_idx]; + const u32 rn_val = state->Reg[rn_idx]; + + // Assume SMMUL by default. + s64 result = (s64)(s32)rn_val * (s64)(s32)rm_val; + + if (ra_idx != 15) { + const u32 ra_val = state->Reg[ra_idx]; + + // SMMLA, otherwise SMMLS + if (BIT(6) == 0) + result += ((s64)ra_val << 32); + else + result = ((s64)ra_val << 32) - result; + } + + if (do_round) + result += 0x80000000; + + state->Reg[rd_idx] = ((result >> 32) & 0xFFFFFFFF); + return 1; + } break; case 0x78: if (BITS(20, 24) == 0x18) diff --git a/src/core/arm/interpreter/armsupp.cpp b/src/core/arm/interpreter/armsupp.cpp index 30519f216..6774f8a74 100644 --- a/src/core/arm/interpreter/armsupp.cpp +++ b/src/core/arm/interpreter/armsupp.cpp @@ -227,8 +227,9 @@ ARMul_CPSRAltered (ARMul_State * state) //state->Cpsr &= ~CBIT; ASSIGNV ((state->Cpsr & VBIT) != 0); //state->Cpsr &= ~VBIT; - ASSIGNS ((state->Cpsr & SBIT) != 0); - //state->Cpsr &= ~SBIT; + ASSIGNQ ((state->Cpsr & QBIT) != 0); + //state->Cpsr &= ~QBIT; + state->GEFlag = (state->Cpsr & 0x000F0000); #ifdef MODET ASSIGNT ((state->Cpsr & TBIT) != 0); //state->Cpsr &= ~TBIT; @@ -443,6 +444,14 @@ ARMul_AddOverflow (ARMul_State * state, ARMword a, ARMword b, ARMword result) ASSIGNV (AddOverflow (a, b, result)); } +/* Assigns the Q flag if the given result is considered an overflow from the addition of a and b */ +void ARMul_AddOverflowQ(ARMul_State* state, ARMword a, ARMword b) +{ + u32 result = a + b; + if (((result ^ a) & (u32)0x80000000) && ((a ^ b) & (u32)0x80000000) == 0) + SETQ; +} + /* Assigns the C flag after an subtraction of a and b to give result. */ void diff --git a/src/core/arm/skyeye_common/armdefs.h b/src/core/arm/skyeye_common/armdefs.h index 28a4a0db4..34eb5aaf7 100644 --- a/src/core/arm/skyeye_common/armdefs.h +++ b/src/core/arm/skyeye_common/armdefs.h @@ -198,7 +198,7 @@ struct ARMul_State //ARMword translate_pc; /* add armv6 flags dyf:2010-08-09 */ - ARMword GEFlag, EFlag, AFlag, QFlags; + ARMword GEFlag, EFlag, AFlag, QFlag; //chy:2003-08-19, used in arm v5e|xscale ARMword SFlag; #ifdef MODET diff --git a/src/core/arm/skyeye_common/armemu.h b/src/core/arm/skyeye_common/armemu.h index 7f7c0e682..3ea14b5a3 100644 --- a/src/core/arm/skyeye_common/armemu.h +++ b/src/core/arm/skyeye_common/armemu.h @@ -34,7 +34,7 @@ #define ZBIT (1L << 30) #define CBIT (1L << 29) #define VBIT (1L << 28) -#define SBIT (1L << 27) +#define QBIT (1L << 27) #define IBIT (1L << 7) #define FBIT (1L << 6) #define IFBITS (3L << 6) @@ -156,13 +156,14 @@ #define R15PCMODE (state->Reg[15] & (R15PCBITS | R15MODEBITS)) #define R15MODE (state->Reg[15] & R15MODEBITS) -#define ECC ((NFLAG << 31) | (ZFLAG << 30) | (CFLAG << 29) | (VFLAG << 28) | (SFLAG << 27)) +#define ECC ((NFLAG << 31) | (ZFLAG << 30) | (CFLAG << 29) | (VFLAG << 28) | (QFLAG << 27)) #define EINT (IFFLAGS << 6) #define ER15INT (IFFLAGS << 26) #define EMODE (state->Mode) +#define EGEBITS (state->GEFlag & 0x000F0000) #ifdef MODET -#define CPSR (ECC | EINT | EMODE | (TFLAG << 5)) +#define CPSR (ECC | EGEBITS | (EFLAG << 9) | (AFLAG << 8) | EINT | (TFLAG << 5) | EMODE) #else #define CPSR (ECC | EINT | EMODE) #endif @@ -601,6 +602,7 @@ extern ARMword ARMul_SwitchMode (ARMul_State *, ARMword, ARMword); extern void ARMul_MSRCpsr (ARMul_State *, ARMword, ARMword); extern void ARMul_SubOverflow (ARMul_State *, ARMword, ARMword, ARMword); extern void ARMul_AddOverflow (ARMul_State *, ARMword, ARMword, ARMword); +extern void ARMul_AddOverflowQ(ARMul_State*, ARMword, ARMword); extern void ARMul_SubCarry (ARMul_State *, ARMword, ARMword, ARMword); extern void ARMul_AddCarry (ARMul_State *, ARMword, ARMword, ARMword); extern tdstate ARMul_ThumbDecode (ARMul_State *, ARMword, ARMword, ARMword *); diff --git a/src/core/hle/service/apt_u.cpp b/src/core/hle/service/apt_u.cpp index e7df72ed7..d8b261ba7 100644 --- a/src/core/hle/service/apt_u.cpp +++ b/src/core/hle/service/apt_u.cpp @@ -315,8 +315,8 @@ Interface::Interface() { if (file.IsOpen()) { // Read shared font data - shared_font.resize(file.GetSize()); - file.ReadBytes(shared_font.data(), file.GetSize()); + shared_font.resize((size_t)file.GetSize()); + file.ReadBytes(shared_font.data(), (size_t)file.GetSize()); // Create shared font memory object shared_font_mem = Kernel::CreateSharedMemory("APT_U:shared_font_mem"); diff --git a/src/core/hle/service/fs/fs_user.cpp b/src/core/hle/service/fs/fs_user.cpp index 027c68d01..b1a465274 100644 --- a/src/core/hle/service/fs/fs_user.cpp +++ b/src/core/hle/service/fs/fs_user.cpp @@ -426,6 +426,23 @@ static void IsSdmcDetected(Service::Interface* self) { } /** + * FS_User::IsSdmcWriteable service function + * Outputs: + * 0 : Command header 0x08180000 + * 1 : Result of function, 0 on success, otherwise error code + * 2 : Whether the Sdmc is currently writeable + */ +static void IsSdmcWriteable(Service::Interface* self) { + u32* cmd_buff = Kernel::GetCommandBuffer(); + + cmd_buff[1] = RESULT_SUCCESS.raw; + // If the SD isn't enabled, it can't be writeable...else, stubbed true + cmd_buff[2] = Settings::values.use_virtual_sd ? 1 : 0; + + LOG_DEBUG(Service_FS, " (STUBBED)"); +} + +/** * FS_User::FormatSaveData service function, * formats the SaveData specified by the input path. * Inputs: @@ -510,7 +527,7 @@ const FSUserInterface::FunctionInfo FunctionTable[] = { {0x08150000, nullptr, "GetNandArchiveResource"}, {0x08160000, nullptr, "GetSdmcFatfsErro"}, {0x08170000, IsSdmcDetected, "IsSdmcDetected"}, - {0x08180000, nullptr, "IsSdmcWritable"}, + {0x08180000, IsSdmcWriteable, "IsSdmcWritable"}, {0x08190042, nullptr, "GetSdmcCid"}, {0x081A0042, nullptr, "GetNandCid"}, {0x081B0000, nullptr, "GetSdmcSpeedInfo"}, diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp index ea29e5027..2083357fe 100644 --- a/src/video_core/command_processor.cpp +++ b/src/video_core/command_processor.cpp @@ -90,7 +90,7 @@ static inline void WritePicaReg(u32 id, u32 value, u32 mask) { const auto& index_info = registers.index_array; const u8* index_address_8 = Memory::GetPointer(PAddrToVAddr(base_address + index_info.offset)); const u16* index_address_16 = (u16*)index_address_8; - bool index_u16 = (bool)index_info.format; + bool index_u16 = index_info.format != 0; DebugUtils::GeometryDumper geometry_dumper; PrimitiveAssembler<VertexShader::OutputVertex> clipper_primitive_assembler(registers.triangle_topology.Value()); @@ -164,7 +164,7 @@ static inline void WritePicaReg(u32 id, u32 value, u32 mask) { case PICA_REG_INDEX(vs_bool_uniforms): for (unsigned i = 0; i < 16; ++i) - VertexShader::GetBoolUniform(i) = (registers.vs_bool_uniforms.Value() & (1 << i)); + VertexShader::GetBoolUniform(i) = (registers.vs_bool_uniforms.Value() & (1 << i)) != 0; break; diff --git a/src/video_core/pica.h b/src/video_core/pica.h index 1fac9ff36..89d97e4e9 100644 --- a/src/video_core/pica.h +++ b/src/video_core/pica.h @@ -201,9 +201,9 @@ struct Regs { }; const std::array<FullTextureConfig, 3> GetTextures() const { return {{ - { static_cast<bool>(texture0_enable), texture0, texture0_format }, - { static_cast<bool>(texture1_enable), texture1, texture1_format }, - { static_cast<bool>(texture2_enable), texture2, texture2_format } + { texture0_enable.ToBool(), texture0, texture0_format }, + { texture1_enable.ToBool(), texture1, texture1_format }, + { texture2_enable.ToBool(), texture2, texture2_format } }}; } @@ -590,11 +590,11 @@ struct Regs { static std::string GetCommandName(int index) { std::map<u32, std::string> map; - Regs regs; #define ADD_FIELD(name) \ do { \ map.insert({PICA_REG_INDEX(name), #name}); \ - for (u32 i = PICA_REG_INDEX(name) + 1; i < PICA_REG_INDEX(name) + sizeof(regs.name) / 4; ++i) \ + /* TODO: change to Regs::name when VS2015 and other compilers support it */ \ + for (u32 i = PICA_REG_INDEX(name) + 1; i < PICA_REG_INDEX(name) + sizeof(Regs().name) / 4; ++i) \ map.insert({i, #name + std::string("+") + std::to_string(i-PICA_REG_INDEX(name))}); \ } while(false) |