/* * Copyright (C) 2008 The Android Open Source Project * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* common includes */ #include "Dalvik.h" #include "interp/InterpDefs.h" #include "mterp/Mterp.h" #include <math.h> // needed for fmod, fmodf #include "mterp/common/FindInterface.h" /* * Configuration defines. These affect the C implementations, i.e. the * portable interpreter(s) and C stubs. * * Some defines are controlled by the Makefile, e.g.: * WITH_INSTR_CHECKS * WITH_TRACKREF_CHECKS * EASY_GDB * NDEBUG * * If THREADED_INTERP is not defined, we use a classic "while true / switch" * interpreter. If it is defined, then the tail end of each instruction * handler fetches the next instruction and jumps directly to the handler. * This increases the size of the "Std" interpreter by about 10%, but * provides a speedup of about the same magnitude. * * There's a "hybrid" approach that uses a goto table instead of a switch * statement, avoiding the "is the opcode in range" tests required for switch. * The performance is close to the threaded version, and without the 10% * size increase, but the benchmark results are off enough that it's not * worth adding as a third option. */ #define THREADED_INTERP /* threaded vs. while-loop interpreter */ #ifdef WITH_INSTR_CHECKS /* instruction-level paranoia (slow!) */ # define CHECK_BRANCH_OFFSETS # define CHECK_REGISTER_INDICES #endif /* * ARM EABI requires 64-bit alignment for access to 64-bit data types. We * can't just use pointers to copy 64-bit values out of our interpreted * register set, because gcc will generate ldrd/strd. * * The __UNION version copies data in and out of a union. The __MEMCPY * version uses a memcpy() call to do the transfer; gcc is smart enough to * not actually call memcpy(). The __UNION version is very bad on ARM; * it only uses one more instruction than __MEMCPY, but for some reason * gcc thinks it needs separate storage for every instance of the union. * On top of that, it feels the need to zero them out at the start of the * method. Net result is we zero out ~700 bytes of stack space at the top * of the interpreter using ARM STM instructions. */ #if defined(__ARM_EABI__) //# define NO_UNALIGN_64__UNION # define NO_UNALIGN_64__MEMCPY #endif //#define LOG_INSTR /* verbose debugging */ /* set and adjust ANDROID_LOG_TAGS='*:i jdwp:i dalvikvm:i dalvikvmi:i' */ /* * Keep a tally of accesses to fields. Currently only works if full DEX * optimization is disabled. */ #ifdef PROFILE_FIELD_ACCESS # define UPDATE_FIELD_GET(_field) { (_field)->gets++; } # define UPDATE_FIELD_PUT(_field) { (_field)->puts++; } #else # define UPDATE_FIELD_GET(_field) ((void)0) # define UPDATE_FIELD_PUT(_field) ((void)0) #endif /* * Export another copy of the PC on every instruction; this is largely * redundant with EXPORT_PC and the debugger code. This value can be * compared against what we have stored on the stack with EXPORT_PC to * help ensure that we aren't missing any export calls. */ #if WITH_EXTRA_GC_CHECKS > 1 # define EXPORT_EXTRA_PC() (self->currentPc2 = pc) #else # define EXPORT_EXTRA_PC() #endif /* * Adjust the program counter. "_offset" is a signed int, in 16-bit units. * * Assumes the existence of "const u2* pc" and "const u2* curMethod->insns". * * We don't advance the program counter until we finish an instruction or * branch, because we do want to have to unroll the PC if there's an * exception. */ #ifdef CHECK_BRANCH_OFFSETS # define ADJUST_PC(_offset) do { \ int myoff = _offset; /* deref only once */ \ if (pc + myoff < curMethod->insns || \ pc + myoff >= curMethod->insns + dvmGetMethodInsnsSize(curMethod)) \ { \ char* desc; \ desc = dexProtoCopyMethodDescriptor(&curMethod->prototype); \ LOGE("Invalid branch %d at 0x%04x in %s.%s %s\n", \ myoff, (int) (pc - curMethod->insns), \ curMethod->clazz->descriptor, curMethod->name, desc); \ free(desc); \ dvmAbort(); \ } \ pc += myoff; \ EXPORT_EXTRA_PC(); \ } while (false) #else # define ADJUST_PC(_offset) do { \ pc += _offset; \ EXPORT_EXTRA_PC(); \ } while (false) #endif /* * If enabled, log instructions as we execute them. */ #ifdef LOG_INSTR # define ILOGD(...) ILOG(LOG_DEBUG, __VA_ARGS__) # define ILOGV(...) ILOG(LOG_VERBOSE, __VA_ARGS__) # define ILOG(_level, ...) do { \ char debugStrBuf[128]; \ snprintf(debugStrBuf, sizeof(debugStrBuf), __VA_ARGS__); \ if (curMethod != NULL) \ LOG(_level, LOG_TAG"i", "%-2d|%04x%s\n", \ self->threadId, (int)(pc - curMethod->insns), debugStrBuf); \ else \ LOG(_level, LOG_TAG"i", "%-2d|####%s\n", \ self->threadId, debugStrBuf); \ } while(false) void dvmDumpRegs(const Method* method, const u4* framePtr, bool inOnly); # define DUMP_REGS(_meth, _frame, _inOnly) dvmDumpRegs(_meth, _frame, _inOnly) static const char kSpacing[] = " "; #else # define ILOGD(...) ((void)0) # define ILOGV(...) ((void)0) # define DUMP_REGS(_meth, _frame, _inOnly) ((void)0) #endif /* get a long from an array of u4 */ static inline s8 getLongFromArray(const u4* ptr, int idx) { #if defined(NO_UNALIGN_64__UNION) union { s8 ll; u4 parts[2]; } conv; ptr += idx; conv.parts[0] = ptr[0]; conv.parts[1] = ptr[1]; return conv.ll; #elif defined(NO_UNALIGN_64__MEMCPY) s8 val; memcpy(&val, &ptr[idx], 8); return val; #else return *((s8*) &ptr[idx]); #endif } /* store a long into an array of u4 */ static inline void putLongToArray(u4* ptr, int idx, s8 val) { #if defined(NO_UNALIGN_64__UNION) union { s8 ll; u4 parts[2]; } conv; ptr += idx; conv.ll = val; ptr[0] = conv.parts[0]; ptr[1] = conv.parts[1]; #elif defined(NO_UNALIGN_64__MEMCPY) memcpy(&ptr[idx], &val, 8); #else *((s8*) &ptr[idx]) = val; #endif } /* get a double from an array of u4 */ static inline double getDoubleFromArray(const u4* ptr, int idx) { #if defined(NO_UNALIGN_64__UNION) union { double d; u4 parts[2]; } conv; ptr += idx; conv.parts[0] = ptr[0]; conv.parts[1] = ptr[1]; return conv.d; #elif defined(NO_UNALIGN_64__MEMCPY) double dval; memcpy(&dval, &ptr[idx], 8); return dval; #else return *((double*) &ptr[idx]); #endif } /* store a double into an array of u4 */ static inline void putDoubleToArray(u4* ptr, int idx, double dval) { #if defined(NO_UNALIGN_64__UNION) union { double d; u4 parts[2]; } conv; ptr += idx; conv.d = dval; ptr[0] = conv.parts[0]; ptr[1] = conv.parts[1]; #elif defined(NO_UNALIGN_64__MEMCPY) memcpy(&ptr[idx], &dval, 8); #else *((double*) &ptr[idx]) = dval; #endif } /* * If enabled, validate the register number on every access. Otherwise, * just do an array access. * * Assumes the existence of "u4* fp". * * "_idx" may be referenced more than once. */ #ifdef CHECK_REGISTER_INDICES # define GET_REGISTER(_idx) \ ( (_idx) < curMethod->registersSize ? \ (fp[(_idx)]) : (assert(!"bad reg"),1969) ) # define SET_REGISTER(_idx, _val) \ ( (_idx) < curMethod->registersSize ? \ (fp[(_idx)] = (u4)(_val)) : (assert(!"bad reg"),1969) ) # define GET_REGISTER_AS_OBJECT(_idx) ((Object *)GET_REGISTER(_idx)) # define SET_REGISTER_AS_OBJECT(_idx, _val) SET_REGISTER(_idx, (s4)_val) # define GET_REGISTER_INT(_idx) ((s4) GET_REGISTER(_idx)) # define SET_REGISTER_INT(_idx, _val) SET_REGISTER(_idx, (s4)_val) # define GET_REGISTER_WIDE(_idx) \ ( (_idx) < curMethod->registersSize-1 ? \ getLongFromArray(fp, (_idx)) : (assert(!"bad reg"),1969) ) # define SET_REGISTER_WIDE(_idx, _val) \ ( (_idx) < curMethod->registersSize-1 ? \ putLongToArray(fp, (_idx), (_val)) : (assert(!"bad reg"),1969) ) # define GET_REGISTER_FLOAT(_idx) \ ( (_idx) < curMethod->registersSize ? \ (*((float*) &fp[(_idx)])) : (assert(!"bad reg"),1969.0f) ) # define SET_REGISTER_FLOAT(_idx, _val) \ ( (_idx) < curMethod->registersSize ? \ (*((float*) &fp[(_idx)]) = (_val)) : (assert(!"bad reg"),1969.0f) ) # define GET_REGISTER_DOUBLE(_idx) \ ( (_idx) < curMethod->registersSize-1 ? \ getDoubleFromArray(fp, (_idx)) : (assert(!"bad reg"),1969.0) ) # define SET_REGISTER_DOUBLE(_idx, _val) \ ( (_idx) < curMethod->registersSize-1 ? \ putDoubleToArray(fp, (_idx), (_val)) : (assert(!"bad reg"),1969.0) ) #else # define GET_REGISTER(_idx) (fp[(_idx)]) # define SET_REGISTER(_idx, _val) (fp[(_idx)] = (_val)) # define GET_REGISTER_AS_OBJECT(_idx) ((Object*) fp[(_idx)]) # define SET_REGISTER_AS_OBJECT(_idx, _val) (fp[(_idx)] = (u4)(_val)) # define GET_REGISTER_INT(_idx) ((s4)GET_REGISTER(_idx)) # define SET_REGISTER_INT(_idx, _val) SET_REGISTER(_idx, (s4)_val) # define GET_REGISTER_WIDE(_idx) getLongFromArray(fp, (_idx)) # define SET_REGISTER_WIDE(_idx, _val) putLongToArray(fp, (_idx), (_val)) # define GET_REGISTER_FLOAT(_idx) (*((float*) &fp[(_idx)])) # define SET_REGISTER_FLOAT(_idx, _val) (*((float*) &fp[(_idx)]) = (_val)) # define GET_REGISTER_DOUBLE(_idx) getDoubleFromArray(fp, (_idx)) # define SET_REGISTER_DOUBLE(_idx, _val) putDoubleToArray(fp, (_idx), (_val)) #endif /* * Get 16 bits from the specified offset of the program counter. We always * want to load 16 bits at a time from the instruction stream -- it's more * efficient than 8 and won't have the alignment problems that 32 might. * * Assumes existence of "const u2* pc". */ #define FETCH(_offset) (pc[(_offset)]) /* * Extract instruction byte from 16-bit fetch (_inst is a u2). */ #define INST_INST(_inst) ((_inst) & 0xff) /* * Replace the opcode (used when handling breakpoints). _opcode is a u1. */ #define INST_REPLACE_OP(_inst, _opcode) (((_inst) & 0xff00) | _opcode) /* * Extract the "vA, vB" 4-bit registers from the instruction word (_inst is u2). */ #define INST_A(_inst) (((_inst) >> 8) & 0x0f) #define INST_B(_inst) ((_inst) >> 12) /* * Get the 8-bit "vAA" 8-bit register index from the instruction word. * (_inst is u2) */ #define INST_AA(_inst) ((_inst) >> 8) /* * The current PC must be available to Throwable constructors, e.g. * those created by dvmThrowException(), so that the exception stack * trace can be generated correctly. If we don't do this, the offset * within the current method won't be shown correctly. See the notes * in Exception.c. * * This is also used to determine the address for precise GC. * * Assumes existence of "u4* fp" and "const u2* pc". */ #define EXPORT_PC() (SAVEAREA_FROM_FP(fp)->xtra.currentPc = pc) /* * Determine if we need to switch to a different interpreter. "_current" * is either INTERP_STD or INTERP_DBG. It should be fixed for a given * interpreter generation file, which should remove the outer conditional * from the following. * * If we're building without debug and profiling support, we never switch. */ #if defined(WITH_JIT) # define NEED_INTERP_SWITCH(_current) ( \ (_current == INTERP_STD) ? \ dvmJitDebuggerOrProfilerActive() : !dvmJitDebuggerOrProfilerActive() ) #else # define NEED_INTERP_SWITCH(_current) ( \ (_current == INTERP_STD) ? \ dvmDebuggerOrProfilerActive() : !dvmDebuggerOrProfilerActive() ) #endif /* * Check to see if "obj" is NULL. If so, throw an exception. Assumes the * pc has already been exported to the stack. * * Perform additional checks on debug builds. * * Use this to check for NULL when the instruction handler calls into * something that could throw an exception (so we have already called * EXPORT_PC at the top). */ static inline bool checkForNull(Object* obj) { if (obj == NULL) { dvmThrowException("Ljava/lang/NullPointerException;", NULL); return false; } #ifdef WITH_EXTRA_OBJECT_VALIDATION if (!dvmIsValidObject(obj)) { LOGE("Invalid object %p\n", obj); dvmAbort(); } #endif #ifndef NDEBUG if (obj->clazz == NULL || ((u4) obj->clazz) <= 65536) { /* probable heap corruption */ LOGE("Invalid object class %p (in %p)\n", obj->clazz, obj); dvmAbort(); } #endif return true; } /* * Check to see if "obj" is NULL. If so, export the PC into the stack * frame and throw an exception. * * Perform additional checks on debug builds. * * Use this to check for NULL when the instruction handler doesn't do * anything else that can throw an exception. */ static inline bool checkForNullExportPC(Object* obj, u4* fp, const u2* pc) { if (obj == NULL) { EXPORT_PC(); dvmThrowException("Ljava/lang/NullPointerException;", NULL); return false; } #ifdef WITH_EXTRA_OBJECT_VALIDATION if (!dvmIsValidObject(obj)) { LOGE("Invalid object %p\n", obj); dvmAbort(); } #endif #ifndef NDEBUG if (obj->clazz == NULL || ((u4) obj->clazz) <= 65536) { /* probable heap corruption */ LOGE("Invalid object class %p (in %p)\n", obj->clazz, obj); dvmAbort(); } #endif return true; }