src/arch/x86/predecoder.cc

   1 /*
   2  * Copyright (c) 2007 The Hewlett-Packard Development Company
   3  * All rights reserved.
   4  *
   5  * Redistribution and use of this software in source and binary forms,
   6  * with or without modification, are permitted provided that the
   7  * following conditions are met:
   8  *
   9  * The software must be used only for Non-Commercial Use which means any
  10  * use which is NOT directed to receiving any direct monetary
  11  * compensation for, or commercial advantage from such use.  Illustrative
  12  * examples of non-commercial use are academic research, personal study,
  13  * teaching, education and corporate research & development.
  14  * Illustrative examples of commercial use are distributing products for
  15  * commercial advantage and providing services using the software for
  16  * commercial advantage.
  17  *
  18  * If you wish to use this software or functionality therein that may be
  19  * covered by patents for commercial use, please contact:
  20  *     Director of Intellectual Property Licensing
  21  *     Office of Strategy and Technology
  22  *     Hewlett-Packard Company
  23  *     1501 Page Mill Road
  24  *     Palo Alto, California  94304
  25  *
  26  * Redistributions of source code must retain the above copyright notice,
  27  * this list of conditions and the following disclaimer.  Redistributions
  28  * in binary form must reproduce the above copyright notice, this list of
  29  * conditions and the following disclaimer in the documentation and/or
  30  * other materials provided with the distribution.  Neither the name of
  31  * the COPYRIGHT HOLDER(s), HEWLETT-PACKARD COMPANY, nor the names of its
  32  * contributors may be used to endorse or promote products derived from
  33  * this software without specific prior written permission.  No right of
  34  * sublicense is granted herewith.  Derivatives of the software and
  35  * output created using the software may be prepared, but only for
  36  * Non-Commercial Uses.  Derivatives of the software may be shared with
  37  * others provided: (i) the others agree to abide by the list of
  38  * conditions herein which includes the Non-Commercial Use restrictions;
  39  * and (ii) such Derivatives of the software include the above copyright
  40  * notice to acknowledge the contribution from this software where
  41  * applicable, this list of conditions and the disclaimer below.
  42  *
  43  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  44  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  45  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
  46  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
  47  * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  48  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
  49  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  50  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  51  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  52  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  53  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  54  *
  55  * Authors: Gabe Black
  56  */
  57
  58 #include "arch/x86/predecoder.hh"
  59 #include "base/misc.hh"
  60 #include "base/trace.hh"
  61 #include "sim/host.hh"
  62
  63 namespace X86ISA
  64 {
  65     void Predecoder::reset()
  66     {
  67         origPC = basePC + offset;
  68         DPRINTF(Predecoder, "Setting origPC to %#x\n", origPC);
  69         emi.rex = 0;
  70         emi.legacy = 0;
  71         emi.opcode.num = 0;
  72
  73         immediateCollected = 0;
  74         emi.immediate = 0;
  75         emi.displacement = 0;
  76
  77         emi.modRM = 0;
  78         emi.sib = 0;
  79         emi.mode = 0;
  80     }
  81
  82     void Predecoder::process()
  83     {
  84         //This function drives the predecoder state machine.
  85
  86         //Some sanity checks. You shouldn't try to process more bytes if
  87         //there aren't any, and you shouldn't overwrite an already
  88         //predecoder ExtMachInst.
  89         assert(!outOfBytes);
  90         assert(!emiIsReady);
  91
  92         //While there's still something to do...
  93         while(!emiIsReady && !outOfBytes)
  94         {
  95             uint8_t nextByte = getNextByte();
  96             switch(state)
  97             {
  98               case ResetState:
  99                 reset();
 100                 state = PrefixState;
 101               case PrefixState:
 102                 state = doPrefixState(nextByte);
 103                 break;
 104               case OpcodeState:
 105                 state = doOpcodeState(nextByte);
 106                 break;
 107               case ModRMState:
 108                 state = doModRMState(nextByte);
 109                 break;
 110               case SIBState:
 111                 state = doSIBState(nextByte);
 112                 break;
 113               case DisplacementState:
 114                 state = doDisplacementState();
 115                 break;
 116               case ImmediateState:
 117                 state = doImmediateState();
 118                 break;
 119               case ErrorState:
 120                 panic("Went to the error state in the predecoder.\n");
 121               default:
 122                 panic("Unrecognized state! %d\n", state);
 123             }
 124         }
 125     }
 126
 127     //Either get a prefix and record it in the ExtMachInst, or send the
 128     //state machine on to get the opcode(s).
 129     Predecoder::State Predecoder::doPrefixState(uint8_t nextByte)
 130     {
 131         uint8_t prefix = Prefixes[nextByte];
 132         State nextState = PrefixState;
 133         if(prefix)
 134             consumeByte();
 135         switch(prefix)
 136         {
 137             //Operand size override prefixes
 138           case OperandSizeOverride:
 139             DPRINTF(Predecoder, "Found operand size override prefix.\n");
 140             emi.legacy.op = true;
 141             break;
 142           case AddressSizeOverride:
 143             DPRINTF(Predecoder, "Found address size override prefix.\n");
 144             emi.legacy.addr = true;
 145             break;
 146             //Segment override prefixes
 147           case CSOverride:
 148           case DSOverride:
 149           case ESOverride:
 150           case FSOverride:
 151           case GSOverride:
 152           case SSOverride:
 153             DPRINTF(Predecoder, "Found segment override.\n");
 154             emi.legacy.seg = prefix;
 155             break;
 156           case Lock:
 157             DPRINTF(Predecoder, "Found lock prefix.\n");
 158             emi.legacy.lock = true;
 159             break;
 160           case Rep:
 161             DPRINTF(Predecoder, "Found rep prefix.\n");
 162             emi.legacy.rep = true;
 163             break;
 164           case Repne:
 165             DPRINTF(Predecoder, "Found repne prefix.\n");
 166             emi.legacy.repne = true;
 167             break;
 168           case RexPrefix:
 169             DPRINTF(Predecoder, "Found Rex prefix %#x.\n", nextByte);
 170             emi.rex = nextByte;
 171             break;
 172           case 0:
 173             nextState = OpcodeState;
 174             break;
 175           default:
 176             panic("Unrecognized prefix %#x\n", nextByte);
 177         }
 178         return nextState;
 179     }
 180
 181     //Load all the opcodes (currently up to 2) and then figure out
 182     //what immediate and/or ModRM is needed.
 183     Predecoder::State Predecoder::doOpcodeState(uint8_t nextByte)
 184     {
 185         State nextState = ErrorState;
 186         emi.opcode.num++;
 187         //We can't handle 3+ byte opcodes right now
 188         assert(emi.opcode.num < 3);
 189         consumeByte();
 190         if(emi.opcode.num == 1 && nextByte == 0x0f)
 191         {
 192             nextState = OpcodeState;
 193             DPRINTF(Predecoder, "Found two byte opcode.\n");
 194             emi.opcode.prefixA = nextByte;
 195         }
 196         else if(emi.opcode.num == 2 &&
 197                 (nextByte == 0x0f ||
 198                  (nextByte & 0xf8) == 0x38))
 199         {
 200             panic("Three byte opcodes aren't yet supported!\n");
 201             nextState = OpcodeState;
 202             DPRINTF(Predecoder, "Found three byte opcode.\n");
 203             emi.opcode.prefixB = nextByte;
 204         }
 205         else
 206         {
 207             DPRINTF(Predecoder, "Found opcode %#x.\n", nextByte);
 208             emi.opcode.op = nextByte;
 209
 210             //Figure out the effective operand size. This can be overriden to
 211             //a fixed value at the decoder level.
 212             int logOpSize;
 213             if(/*FIXME long mode*/1)
 214             {
 215                 if(emi.rex.w)
 216                     logOpSize = 3; // 64 bit operand size
 217                 else if(emi.legacy.op)
 218                     logOpSize = 1; // 16 bit operand size
 219                 else
 220                     logOpSize = 2; // 32 bit operand size
 221             }
 222             else if(/*FIXME default 32*/1)
 223             {
 224                 if(emi.legacy.op)
 225                     logOpSize = 1; // 16 bit operand size
 226                 else
 227                     logOpSize = 2; // 32 bit operand size
 228             }
 229             else // 16 bit default operand size
 230             {
 231                 if(emi.legacy.op)
 232                     logOpSize = 2; // 32 bit operand size
 233                 else
 234                     logOpSize = 1; // 16 bit operand size
 235             }
 236
 237             //Set the actual op size
 238             emi.opSize = 1 << logOpSize;
 239
 240             //Figure out the effective address size. This can be overriden to
 241             //a fixed value at the decoder level.
 242             int logAddrSize;
 243             if(/*FIXME 64-bit mode*/1)
 244             {
 245                 if(emi.legacy.addr)
 246                     logAddrSize = 2; // 32 bit address size
 247                 else
 248                     logAddrSize = 3; // 64 bit address size
 249             }
 250             else if(/*FIXME default 32*/1)
 251             {
 252                 if(emi.legacy.addr)
 253                     logAddrSize = 1; // 16 bit address size
 254                 else
 255                     logAddrSize = 2; // 32 bit address size
 256             }
 257             else // 16 bit default operand size
 258             {
 259                 if(emi.legacy.addr)
 260                     logAddrSize = 2; // 32 bit address size
 261                 else
 262                     logAddrSize = 1; // 16 bit address size
 263             }
 264
 265             //Set the actual address size
 266             emi.addrSize = 1 << logAddrSize;
 267
 268             //Figure out how big of an immediate we'll retreive based
 269             //on the opcode.
 270             int immType = ImmediateType[emi.opcode.num - 1][nextByte];
 271             immediateSize = SizeTypeToSize[logOpSize - 1][immType];
 272
 273             //Determine what to expect next
 274             if (UsesModRM[emi.opcode.num - 1][nextByte]) {
 275                 nextState = ModRMState;
 276             } else {
 277                 if(immediateSize) {
 278                     nextState = ImmediateState;
 279                 } else {
 280                     emiIsReady = true;
 281                     nextState = ResetState;
 282                 }
 283             }
 284         }
 285         return nextState;
 286     }
 287
 288     //Get the ModRM byte and determine what displacement, if any, there is.
 289     //Also determine whether or not to get the SIB byte, displacement, or
 290     //immediate next.
 291     Predecoder::State Predecoder::doModRMState(uint8_t nextByte)
 292     {
 293         State nextState = ErrorState;
 294         ModRM modRM;
 295         modRM = nextByte;
 296         DPRINTF(Predecoder, "Found modrm byte %#x.\n", nextByte);
 297         if (0) {//FIXME in 16 bit mode
 298             //figure out 16 bit displacement size
 299             if(modRM.mod == 0 && modRM.rm == 6 || modRM.mod == 2)
 300                 displacementSize = 2;
 301             else if(modRM.mod == 1)
 302                 displacementSize = 1;
 303             else
 304                 displacementSize = 0;
 305         } else {
 306             //figure out 32/64 bit displacement size
 307             if(modRM.mod == 0 && modRM.rm == 5 || modRM.mod == 2)
 308                 displacementSize = 4;
 309             else if(modRM.mod == 1)
 310                 displacementSize = 1;
 311             else
 312                 displacementSize = 0;
 313         }
 314
 315         // The "test" instruction in group 3 needs an immediate, even though
 316         // the other instructions with the same actual opcode don't.
 317         if (emi.opcode.num == 1 && (modRM.reg & 0x6) == 0) {
 318            if (emi.opcode.op == 0xF6)
 319                immediateSize = 1;
 320            else if (emi.opcode.op == 0xF7)
 321                immediateSize = (emi.opSize == 8) ? 4 : emi.opSize;
 322         }
 323
 324         //If there's an SIB, get that next.
 325         //There is no SIB in 16 bit mode.
 326         if (modRM.rm == 4 && modRM.mod != 3) {
 327                 // && in 32/64 bit mode)
 328             nextState = SIBState;
 329         } else if(displacementSize) {
 330             nextState = DisplacementState;
 331         } else if(immediateSize) {
 332             nextState = ImmediateState;
 333         } else {
 334             emiIsReady = true;
 335             nextState = ResetState;
 336         }
 337         //The ModRM byte is consumed no matter what
 338         consumeByte();
 339         emi.modRM = modRM;
 340         return nextState;
 341     }
 342
 343     //Get the SIB byte. We don't do anything with it at this point, other
 344     //than storing it in the ExtMachInst. Determine if we need to get a
 345     //displacement or immediate next.
 346     Predecoder::State Predecoder::doSIBState(uint8_t nextByte)
 347     {
 348         State nextState = ErrorState;
 349         emi.sib = nextByte;
 350         DPRINTF(Predecoder, "Found SIB byte %#x.\n", nextByte);
 351         consumeByte();
 352         if (emi.modRM.mod == 0 && emi.sib.base == 5)
 353             displacementSize = 4;
 354         if (displacementSize) {
 355             nextState = DisplacementState;
 356         } else if(immediateSize) {
 357             nextState = ImmediateState;
 358         } else {
 359             emiIsReady = true;
 360             nextState = ResetState;
 361         }
 362         return nextState;
 363     }
 364
 365     //Gather up the displacement, or at least as much of it
 366     //as we can get.
 367     Predecoder::State Predecoder::doDisplacementState()
 368     {
 369         State nextState = ErrorState;
 370
 371         getImmediate(immediateCollected,
 372                 emi.displacement,
 373                 displacementSize);
 374
 375         DPRINTF(Predecoder, "Collecting %d byte displacement, got %d bytes.\n",
 376                 displacementSize, immediateCollected);
 377
 378         if(displacementSize == immediateCollected) {
 379             //Reset this for other immediates.
 380             immediateCollected = 0;
 381             //Sign extend the displacement
 382             switch(displacementSize)
 383             {
 384               case 1:
 385                 emi.displacement = sext<8>(emi.displacement);
 386                 break;
 387               case 2:
 388                 emi.displacement = sext<16>(emi.displacement);
 389                 break;
 390               case 4:
 391                 emi.displacement = sext<32>(emi.displacement);
 392                 break;
 393               default:
 394                 panic("Undefined displacement size!\n");
 395             }
 396             DPRINTF(Predecoder, "Collected displacement %#x.\n",
 397                     emi.displacement);
 398             if(immediateSize) {
 399                 nextState = ImmediateState;
 400             } else {
 401                 emiIsReady = true;
 402                 nextState = ResetState;
 403             }
 404         }
 405         else
 406             nextState = DisplacementState;
 407         return nextState;
 408     }
 409
 410     //Gather up the immediate, or at least as much of it
 411     //as we can get
 412     Predecoder::State Predecoder::doImmediateState()
 413     {
 414         State nextState = ErrorState;
 415
 416         getImmediate(immediateCollected,
 417                 emi.immediate,
 418                 immediateSize);
 419
 420         DPRINTF(Predecoder, "Collecting %d byte immediate, got %d bytes.\n",
 421                 immediateSize, immediateCollected);
 422
 423         if(immediateSize == immediateCollected)
 424         {
 425             //Reset this for other immediates.
 426             immediateCollected = 0;
 427
 428             //XXX Warning! The following is an observed pattern and might
 429             //not always be true!
 430
 431             //Instructions which use 64 bit operands but 32 bit immediates
 432             //need to have the immediate sign extended to 64 bits.
 433             //Instructions which use true 64 bit immediates won't be
 434             //affected, and instructions that use true 32 bit immediates
 435             //won't notice.
 436             switch(immediateSize)
 437             {
 438               case 4:
 439                 emi.immediate = sext<32>(emi.immediate);
 440                 break;
 441               case 1:
 442                 emi.immediate = sext<8>(emi.immediate);
 443             }
 444
 445             DPRINTF(Predecoder, "Collected immediate %#x.\n",
 446                     emi.immediate);
 447             emiIsReady = true;
 448             nextState = ResetState;
 449         }
 450         else
 451             nextState = ImmediateState;
 452         return nextState;
 453     }
 454 }