src/arch/x86/predecoder.cc

   1 /*
   2  * Copyright (c) 2007-2008 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/miscregs.hh"
  59 #include "arch/x86/predecoder.hh"
  60 #include "base/misc.hh"
  61 #include "base/trace.hh"
  62 #include "base/types.hh"
  63 #include "cpu/thread_context.hh"
  64
  65 namespace X86ISA
  66 {
  67     void Predecoder::doReset()
  68     {
  69         origPC = basePC + offset;
  70         DPRINTF(Predecoder, "Setting origPC to %#x\n", origPC);
  71         emi.rex = 0;
  72         emi.legacy = 0;
  73         emi.opcode.num = 0;
  74         emi.opcode.op = 0;
  75         emi.opcode.prefixA = emi.opcode.prefixB = 0;
  76
  77         immediateCollected = 0;
  78         emi.immediate = 0;
  79         emi.displacement = 0;
  80         emi.dispSize = 0;
  81
  82         emi.modRM = 0;
  83         emi.sib = 0;
  84         m5Reg = tc->readMiscRegNoEffect(MISCREG_M5_REG);
  85         emi.mode.mode = m5Reg.mode;
  86         emi.mode.submode = m5Reg.submode;
  87     }
  88
  89     void Predecoder::process()
  90     {
  91         //This function drives the predecoder state machine.
  92
  93         //Some sanity checks. You shouldn't try to process more bytes if
  94         //there aren't any, and you shouldn't overwrite an already
  95         //predecoder ExtMachInst.
  96         assert(!outOfBytes);
  97         assert(!emiIsReady);
  98
  99         //While there's still something to do...
 100         while(!emiIsReady && !outOfBytes)
 101         {
 102             uint8_t nextByte = getNextByte();
 103             switch(state)
 104             {
 105               case ResetState:
 106                 doReset();
 107                 state = PrefixState;
 108               case PrefixState:
 109                 state = doPrefixState(nextByte);
 110                 break;
 111               case OpcodeState:
 112                 state = doOpcodeState(nextByte);
 113                 break;
 114               case ModRMState:
 115                 state = doModRMState(nextByte);
 116                 break;
 117               case SIBState:
 118                 state = doSIBState(nextByte);
 119                 break;
 120               case DisplacementState:
 121                 state = doDisplacementState();
 122                 break;
 123               case ImmediateState:
 124                 state = doImmediateState();
 125                 break;
 126               case ErrorState:
 127                 panic("Went to the error state in the predecoder.\n");
 128               default:
 129                 panic("Unrecognized state! %d\n", state);
 130             }
 131         }
 132     }
 133
 134     //Either get a prefix and record it in the ExtMachInst, or send the
 135     //state machine on to get the opcode(s).
 136     Predecoder::State Predecoder::doPrefixState(uint8_t nextByte)
 137     {
 138         uint8_t prefix = Prefixes[nextByte];
 139         State nextState = PrefixState;
 140         // REX prefixes are only recognized in 64 bit mode.
 141         if (prefix == RexPrefix && emi.mode.submode != SixtyFourBitMode)
 142             prefix = 0;
 143         if (prefix)
 144             consumeByte();
 145         switch(prefix)
 146         {
 147             //Operand size override prefixes
 148           case OperandSizeOverride:
 149             DPRINTF(Predecoder, "Found operand size override prefix.\n");
 150             emi.legacy.op = true;
 151             break;
 152           case AddressSizeOverride:
 153             DPRINTF(Predecoder, "Found address size override prefix.\n");
 154             emi.legacy.addr = true;
 155             break;
 156             //Segment override prefixes
 157           case CSOverride:
 158           case DSOverride:
 159           case ESOverride:
 160           case FSOverride:
 161           case GSOverride:
 162           case SSOverride:
 163             DPRINTF(Predecoder, "Found segment override.\n");
 164             emi.legacy.seg = prefix;
 165             break;
 166           case Lock:
 167             DPRINTF(Predecoder, "Found lock prefix.\n");
 168             emi.legacy.lock = true;
 169             break;
 170           case Rep:
 171             DPRINTF(Predecoder, "Found rep prefix.\n");
 172             emi.legacy.rep = true;
 173             break;
 174           case Repne:
 175             DPRINTF(Predecoder, "Found repne prefix.\n");
 176             emi.legacy.repne = true;
 177             break;
 178           case RexPrefix:
 179             DPRINTF(Predecoder, "Found Rex prefix %#x.\n", nextByte);
 180             emi.rex = nextByte;
 181             break;
 182           case 0:
 183             nextState = OpcodeState;
 184             break;
 185           default:
 186             panic("Unrecognized prefix %#x\n", nextByte);
 187         }
 188         return nextState;
 189     }
 190
 191     //Load all the opcodes (currently up to 2) and then figure out
 192     //what immediate and/or ModRM is needed.
 193     Predecoder::State Predecoder::doOpcodeState(uint8_t nextByte)
 194     {
 195         State nextState = ErrorState;
 196         emi.opcode.num++;
 197         //We can't handle 3+ byte opcodes right now
 198         assert(emi.opcode.num < 4);
 199         consumeByte();
 200         if(emi.opcode.num == 1 && nextByte == 0x0f)
 201         {
 202             nextState = OpcodeState;
 203             DPRINTF(Predecoder, "Found two byte opcode.\n");
 204             emi.opcode.prefixA = nextByte;
 205         }
 206         else if(emi.opcode.num == 2 && (nextByte == 0x38 || nextByte == 0x3F))
 207         {
 208             nextState = OpcodeState;
 209             DPRINTF(Predecoder, "Found three byte opcode.\n");
 210             emi.opcode.prefixB = nextByte;
 211         }
 212         else
 213         {
 214             DPRINTF(Predecoder, "Found opcode %#x.\n", nextByte);
 215             emi.opcode.op = nextByte;
 216
 217             //Figure out the effective operand size. This can be overriden to
 218             //a fixed value at the decoder level.
 219             int logOpSize;
 220             if (emi.rex.w)
 221                 logOpSize = 3; // 64 bit operand size
 222             else if (emi.legacy.op)
 223                 logOpSize = m5Reg.altOp;
 224             else
 225                 logOpSize = m5Reg.defOp;
 226
 227             //Set the actual op size
 228             emi.opSize = 1 << logOpSize;
 229
 230             //Figure out the effective address size. This can be overriden to
 231             //a fixed value at the decoder level.
 232             int logAddrSize;
 233             if(emi.legacy.addr)
 234                 logAddrSize = m5Reg.altAddr;
 235             else
 236                 logAddrSize = m5Reg.defAddr;
 237
 238             //Set the actual address size
 239             emi.addrSize = 1 << logAddrSize;
 240
 241             //Figure out the effective stack width. This can be overriden to
 242             //a fixed value at the decoder level.
 243             emi.stackSize = 1 << m5Reg.stack;
 244
 245             //Figure out how big of an immediate we'll retreive based
 246             //on the opcode.
 247             int immType = ImmediateType[emi.opcode.num - 1][nextByte];
 248             if (emi.opcode.num == 1 && nextByte >= 0xA0 && nextByte <= 0xA3)
 249                 immediateSize = SizeTypeToSize[logAddrSize - 1][immType];
 250             else
 251                 immediateSize = SizeTypeToSize[logOpSize - 1][immType];
 252
 253             //Determine what to expect next
 254             if (UsesModRM[emi.opcode.num - 1][nextByte]) {
 255                 nextState = ModRMState;
 256             } else {
 257                 if(immediateSize) {
 258                     nextState = ImmediateState;
 259                 } else {
 260                     emiIsReady = true;
 261                     nextState = ResetState;
 262                 }
 263             }
 264         }
 265         return nextState;
 266     }
 267
 268     //Get the ModRM byte and determine what displacement, if any, there is.
 269     //Also determine whether or not to get the SIB byte, displacement, or
 270     //immediate next.
 271     Predecoder::State Predecoder::doModRMState(uint8_t nextByte)
 272     {
 273         State nextState = ErrorState;
 274         ModRM modRM;
 275         modRM = nextByte;
 276         DPRINTF(Predecoder, "Found modrm byte %#x.\n", nextByte);
 277         if (m5Reg.defOp == 1) {
 278             //figure out 16 bit displacement size
 279             if ((modRM.mod == 0 && modRM.rm == 6) || modRM.mod == 2)
 280                 displacementSize = 2;
 281             else if (modRM.mod == 1)
 282                 displacementSize = 1;
 283             else
 284                 displacementSize = 0;
 285         } else {
 286             //figure out 32/64 bit displacement size
 287             if ((modRM.mod == 0 && modRM.rm == 5) || modRM.mod == 2)
 288                 displacementSize = 4;
 289             else if (modRM.mod == 1)
 290                 displacementSize = 1;
 291             else
 292                 displacementSize = 0;
 293         }
 294
 295         // The "test" instruction in group 3 needs an immediate, even though
 296         // the other instructions with the same actual opcode don't.
 297         if (emi.opcode.num == 1 && (modRM.reg & 0x6) == 0) {
 298            if (emi.opcode.op == 0xF6)
 299                immediateSize = 1;
 300            else if (emi.opcode.op == 0xF7)
 301                immediateSize = (emi.opSize == 8) ? 4 : emi.opSize;
 302         }
 303
 304         //If there's an SIB, get that next.
 305         //There is no SIB in 16 bit mode.
 306         if (modRM.rm == 4 && modRM.mod != 3) {
 307                 // && in 32/64 bit mode)
 308             nextState = SIBState;
 309         } else if(displacementSize) {
 310             nextState = DisplacementState;
 311         } else if(immediateSize) {
 312             nextState = ImmediateState;
 313         } else {
 314             emiIsReady = true;
 315             nextState = ResetState;
 316         }
 317         //The ModRM byte is consumed no matter what
 318         consumeByte();
 319         emi.modRM = modRM;
 320         return nextState;
 321     }
 322
 323     //Get the SIB byte. We don't do anything with it at this point, other
 324     //than storing it in the ExtMachInst. Determine if we need to get a
 325     //displacement or immediate next.
 326     Predecoder::State Predecoder::doSIBState(uint8_t nextByte)
 327     {
 328         State nextState = ErrorState;
 329         emi.sib = nextByte;
 330         DPRINTF(Predecoder, "Found SIB byte %#x.\n", nextByte);
 331         consumeByte();
 332         if (emi.modRM.mod == 0 && emi.sib.base == 5)
 333             displacementSize = 4;
 334         if (displacementSize) {
 335             nextState = DisplacementState;
 336         } else if(immediateSize) {
 337             nextState = ImmediateState;
 338         } else {
 339             emiIsReady = true;
 340             nextState = ResetState;
 341         }
 342         return nextState;
 343     }
 344
 345     //Gather up the displacement, or at least as much of it
 346     //as we can get.
 347     Predecoder::State Predecoder::doDisplacementState()
 348     {
 349         State nextState = ErrorState;
 350
 351         getImmediate(immediateCollected,
 352                 emi.displacement,
 353                 displacementSize);
 354
 355         DPRINTF(Predecoder, "Collecting %d byte displacement, got %d bytes.\n",
 356                 displacementSize, immediateCollected);
 357
 358         if(displacementSize == immediateCollected) {
 359             //Reset this for other immediates.
 360             immediateCollected = 0;
 361             //Sign extend the displacement
 362             switch(displacementSize)
 363             {
 364               case 1:
 365                 emi.displacement = sext<8>(emi.displacement);
 366                 break;
 367               case 2:
 368                 emi.displacement = sext<16>(emi.displacement);
 369                 break;
 370               case 4:
 371                 emi.displacement = sext<32>(emi.displacement);
 372                 break;
 373               default:
 374                 panic("Undefined displacement size!\n");
 375             }
 376             DPRINTF(Predecoder, "Collected displacement %#x.\n",
 377                     emi.displacement);
 378             if(immediateSize) {
 379                 nextState = ImmediateState;
 380             } else {
 381                 emiIsReady = true;
 382                 nextState = ResetState;
 383             }
 384
 385             emi.dispSize = displacementSize;
 386         }
 387         else
 388             nextState = DisplacementState;
 389         return nextState;
 390     }
 391
 392     //Gather up the immediate, or at least as much of it
 393     //as we can get
 394     Predecoder::State Predecoder::doImmediateState()
 395     {
 396         State nextState = ErrorState;
 397
 398         getImmediate(immediateCollected,
 399                 emi.immediate,
 400                 immediateSize);
 401
 402         DPRINTF(Predecoder, "Collecting %d byte immediate, got %d bytes.\n",
 403                 immediateSize, immediateCollected);
 404
 405         if(immediateSize == immediateCollected)
 406         {
 407             //Reset this for other immediates.
 408             immediateCollected = 0;
 409
 410             //XXX Warning! The following is an observed pattern and might
 411             //not always be true!
 412
 413             //Instructions which use 64 bit operands but 32 bit immediates
 414             //need to have the immediate sign extended to 64 bits.
 415             //Instructions which use true 64 bit immediates won't be
 416             //affected, and instructions that use true 32 bit immediates
 417             //won't notice.
 418             switch(immediateSize)
 419             {
 420               case 4:
 421                 emi.immediate = sext<32>(emi.immediate);
 422                 break;
 423               case 1:
 424                 emi.immediate = sext<8>(emi.immediate);
 425             }
 426
 427             DPRINTF(Predecoder, "Collected immediate %#x.\n",
 428                     emi.immediate);
 429             emiIsReady = true;
 430             nextState = ResetState;
 431         }
 432         else
 433             nextState = ImmediateState;
 434         return nextState;
 435     }
 436 }