///////////////////////////////////////////////////////////////////////// // $Id: shift16.cc,v 1.51 2009-10-13 15:28:16 sshwarts Exp $ ///////////////////////////////////////////////////////////////////////// // // Copyright (C) 2001 MandrakeSoft S.A. // // MandrakeSoft S.A. // 43, rue d'Aboukir // 75002 Paris - France // http://www.linux-mandrake.com/ // http://www.mandrakesoft.com/ // // This library is free software; you can redistribute it and/or // modify it under the terms of the GNU Lesser General Public // License as published by the Free Software Foundation; either // version 2 of the License, or (at your option) any later version. // // This library is distributed in the hope that it will be useful, // but WITHOUT ANY WARRANTY; without even the implied warranty of // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU // Lesser General Public License for more details. // // You should have received a copy of the GNU Lesser General Public // License along with this library; if not, write to the Free Software // Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA B 02110-1301 USA ///////////////////////////////////////////////////////////////////////// #define NEED_CPU_REG_SHORTCUTS 1 #include "bochs.h" #include "cpu.h" #define LOG_THIS BX_CPU_THIS_PTR void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHLD_EwGwM(bxInstruction_c *i) { Bit16u op1_16, op2_16, result_16; Bit32u temp_32, result_32; unsigned count; unsigned of, cf; /* op1:op2 << count. result stored in op1 */ if (i->b1() == 0xa4) // 0x1a4 count = i->Ib(); else // 0x1a5 count = CL; count &= 0x1f; // use only 5 LSB's bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); if (!count) return; op2_16 = BX_READ_16BIT_REG(i->nnn()); /* count < 32, since only lower 5 bits used */ temp_32 = ((Bit32u)(op1_16) << 16) | (op2_16); // double formed by op1:op2 result_32 = temp_32 << count; // hack to act like x86 SHLD when count > 16 if (count > 16) { // when count > 16 actually shifting op1:op2:op2 << count, // it is the same as shifting op2:op2 by count-16 result_32 |= (op1_16 << (count - 16)); } result_16 = (Bit16u)(result_32 >> 16); write_RMW_virtual_word(result_16); SET_FLAGS_OSZAPC_LOGIC_16(result_16); cf = (temp_32 >> (32 - count)) & 0x1; of = cf ^ (result_16 >> 15); // of = cf ^ result15 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHLD_EwGwR(bxInstruction_c *i) { Bit16u op1_16, op2_16, result_16; Bit32u temp_32, result_32; unsigned count; unsigned of, cf; /* op1:op2 << count. result stored in op1 */ if (i->b1() == 0xa4) // 0x1a4 count = i->Ib(); else // 0x1a5 count = CL; count &= 0x1f; // use only 5 LSB's if (!count) return; op1_16 = BX_READ_16BIT_REG(i->rm()); op2_16 = BX_READ_16BIT_REG(i->nnn()); /* count < 32, since only lower 5 bits used */ temp_32 = ((Bit32u)(op1_16) << 16) | (op2_16); // double formed by op1:op2 result_32 = temp_32 << count; // hack to act like x86 SHLD when count > 16 if (count > 16) { // when count > 16 actually shifting op1:op2:op2 << count, // it is the same as shifting op2:op2 by count-16 result_32 |= (op1_16 << (count - 16)); } result_16 = (Bit16u)(result_32 >> 16); BX_WRITE_16BIT_REG(i->rm(), result_16); SET_FLAGS_OSZAPC_LOGIC_16(result_16); cf = (temp_32 >> (32 - count)) & 0x1; of = cf ^ (result_16 >> 15); // of = cf ^ result15 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHRD_EwGwM(bxInstruction_c *i) { Bit16u op1_16, op2_16, result_16; Bit32u temp_32, result_32; unsigned count; unsigned cf, of; if (i->b1() == 0xac) // 0x1ac count = i->Ib(); else // 0x1ad count = CL; count &= 0x1f; /* use only 5 LSB's */ bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); if (!count) return; op2_16 = BX_READ_16BIT_REG(i->nnn()); /* count < 32, since only lower 5 bits used */ temp_32 = (op2_16 << 16) | op1_16; // double formed by op2:op1 result_32 = temp_32 >> count; // hack to act like x86 SHRD when count > 16 if (count > 16) { // when count > 16 actually shifting op2:op2:op1 >> count, // it is the same as shifting op2:op2 by count-16 result_32 |= (op1_16 << (32 - count)); } result_16 = (Bit16u) result_32; write_RMW_virtual_word(result_16); SET_FLAGS_OSZAPC_LOGIC_16(result_16); cf = (op1_16 >> (count - 1)) & 0x1; of = (((result_16 << 1) ^ result_16) >> 15) & 0x1; // of = result14 ^ result15 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHRD_EwGwR(bxInstruction_c *i) { Bit16u op1_16, op2_16, result_16; Bit32u temp_32, result_32; unsigned count; unsigned cf, of; if (i->b1() == 0xac) // 0x1ac count = i->Ib(); else // 0x1ad count = CL; count &= 0x1f; /* use only 5 LSB's */ if (!count) return; op1_16 = BX_READ_16BIT_REG(i->rm()); op2_16 = BX_READ_16BIT_REG(i->nnn()); /* count < 32, since only lower 5 bits used */ temp_32 = (op2_16 << 16) | op1_16; // double formed by op2:op1 result_32 = temp_32 >> count; // hack to act like x86 SHRD when count > 16 if (count > 16) { // when count > 16 actually shifting op2:op2:op1 >> count, // it is the same as shifting op2:op2 by count-16 result_32 |= (op1_16 << (32 - count)); } result_16 = (Bit16u) result_32; BX_WRITE_16BIT_REG(i->rm(), result_16); SET_FLAGS_OSZAPC_LOGIC_16(result_16); cf = (op1_16 >> (count - 1)) & 0x1; of = (((result_16 << 1) ^ result_16) >> 15) & 0x1; // of = result14 ^ result15 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::ROL_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned bit0, bit15; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if ((count & 0x0f) == 0) { if (count & 0x10) { bit0 = (op1_16 & 0x1); bit15 = (op1_16 >> 15); // of = cf ^ result15 SET_FLAGS_OxxxxC(bit0 ^ bit15, bit0); } return; } count &= 0x0f; // only use bottom 4 bits result_16 = (op1_16 << count) | (op1_16 >> (16 - count)); /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } bit0 = (result_16 & 0x1); bit15 = (result_16 >> 15); // of = cf ^ result15 SET_FLAGS_OxxxxC(bit0 ^ bit15, bit0); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::ROR_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned bit14, bit15; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if ((count & 0x0f) == 0) { if (count & 0x10) { bit14 = (op1_16 >> 14) & 1; bit15 = (op1_16 >> 15) & 1; // of = result14 ^ result15 SET_FLAGS_OxxxxC(bit14 ^ bit15, bit15); } return; } count &= 0x0f; // use only 4 LSB's result_16 = (op1_16 >> count) | (op1_16 << (16 - count)); /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } bit14 = (result_16 >> 14) & 1; bit15 = (result_16 >> 15) & 1; // of = result14 ^ result15 SET_FLAGS_OxxxxC(bit14 ^ bit15, bit15); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::RCL_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned of, cf; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); count = (count & 0x1f) % 17; /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if (!count) return; if (count==1) { result_16 = (op1_16 << 1) | getB_CF(); } else if (count==16) { result_16 = (getB_CF() << 15) | (op1_16 >> 1); } else { // 2..15 result_16 = (op1_16 << count) | (getB_CF() << (count - 1)) | (op1_16 >> (17 - count)); } /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } cf = (op1_16 >> (16 - count)) & 0x1; of = cf ^ (result_16 >> 15); // of = cf ^ result15 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::RCR_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned of, cf; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); count = (count & 0x1f) % 17; /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if (! count) return; result_16 = (op1_16 >> count) | (getB_CF() << (16 - count)) | (op1_16 << (17 - count)); /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } cf = (op1_16 >> (count - 1)) & 0x1; of = (((result_16 << 1) ^ result_16) >> 15) & 0x1; // of = result15 ^ result14 SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHL_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned of = 0, cf = 0; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); count &= 0x1f; /* use only 5 LSB's */ /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if (!count) return; if (count <= 16) { result_16 = (op1_16 << count); cf = (op1_16 >> (16 - count)) & 0x1; of = cf ^ (result_16 >> 15); // of = cf ^ result15 } else { result_16 = 0; } /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } SET_FLAGS_OSZAPC_LOGIC_16(result_16); SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SHR_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count; unsigned of, cf; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); count &= 0x1f; /* use only 5 LSB's */ /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if (!count) return; result_16 = (op1_16 >> count); /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } cf = (op1_16 >> (count - 1)) & 0x1; // note, that of == result15 if count == 1 and // of == 0 if count >= 2 of = (((result_16 << 1) ^ result_16) >> 15) & 0x1; SET_FLAGS_OSZAPC_LOGIC_16(result_16); SET_FLAGS_OxxxxC(of, cf); } void BX_CPP_AttrRegparmN(1) BX_CPU_C::SAR_Ew(bxInstruction_c *i) { Bit16u op1_16, result_16; unsigned count, cf; if (i->b1() == 0xd3) count = CL; else // 0xc1 or 0xd1 count = i->Ib(); count &= 0x1f; /* use only 5 LSB's */ /* op1 is a register or memory reference */ if (i->modC0()) { op1_16 = BX_READ_16BIT_REG(i->rm()); } else { bx_address eaddr = BX_CPU_CALL_METHODR(i->ResolveModrm, (i)); /* pointer, segment address pair */ op1_16 = read_RMW_virtual_word(i->seg(), eaddr); } if (!count) return; if (count < 16) { if (op1_16 & 0x8000) { result_16 = (op1_16 >> count) | (0xffff << (16 - count)); } else { result_16 = (op1_16 >> count); } cf = (op1_16 >> (count - 1)) & 0x1; } else { if (op1_16 & 0x8000) { result_16 = 0xffff; } else { result_16 = 0; } cf = (result_16 & 0x1); } SET_FLAGS_OSZAPC_LOGIC_16(result_16); /* signed overflow cannot happen in SAR instruction */ SET_FLAGS_OxxxxC(0, cf); /* now write result back to destination */ if (i->modC0()) { BX_WRITE_16BIT_REG(i->rm(), result_16); } else { write_RMW_virtual_word(result_16); } }