/* * Tiny Code Generator for QEMU * * Copyright (c) 2008 Fabrice Bellard * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to deal * in the Software without restriction, including without limitation the rights * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell * copies of the Software, and to permit persons to whom the Software is * furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice shall be included in * all copies or substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN * THE SOFTWARE. */ #include "qemu/osdep.h" #include "qemu-common.h" #include "cpu.h" #include "exec/exec-all.h" #include "tcg.h" #include "tcg-op.h" #include "tcg-mo.h" /* Unicorn: Necessary for check_exit_request */ #include "exec/gen-icount.h" /* Reduce the number of ifdefs below. This assumes that all uses of TCGV_HIGH and TCGV_LOW are properly protected by a conditional that the compiler can eliminate. */ #if TCG_TARGET_REG_BITS == 64 extern TCGv_i32 TCGV_LOW_link_error(TCGContext *, TCGv_i64); extern TCGv_i32 TCGV_HIGH_link_error(TCGContext *, TCGv_i64); #define TCGV_LOW TCGV_LOW_link_error #define TCGV_HIGH TCGV_HIGH_link_error #endif void tcg_gen_op1(TCGContext *ctx, TCGOpcode opc, TCGArg a1) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; } void tcg_gen_op2(TCGContext *ctx, TCGOpcode opc, TCGArg a1, TCGArg a2) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; op->args[1] = a2; } void tcg_gen_op3(TCGContext *ctx, TCGOpcode opc, TCGArg a1, TCGArg a2, TCGArg a3) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; op->args[1] = a2; op->args[2] = a3; } void tcg_gen_op4(TCGContext *ctx, TCGOpcode opc, TCGArg a1, TCGArg a2, TCGArg a3, TCGArg a4) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; op->args[1] = a2; op->args[2] = a3; op->args[3] = a4; } void tcg_gen_op5(TCGContext *ctx, TCGOpcode opc, TCGArg a1, TCGArg a2, TCGArg a3, TCGArg a4, TCGArg a5) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; op->args[1] = a2; op->args[2] = a3; op->args[3] = a4; op->args[4] = a5; } void tcg_gen_op6(TCGContext *ctx, TCGOpcode opc, TCGArg a1, TCGArg a2, TCGArg a3, TCGArg a4, TCGArg a5, TCGArg a6) { TCGOp *op = tcg_emit_op(ctx, opc); op->args[0] = a1; op->args[1] = a2; op->args[2] = a3; op->args[3] = a4; op->args[4] = a5; op->args[5] = a6; } void tcg_gen_mb(TCGContext *ctx, TCGBar mb_type) { if (ctx->tb_cflags & CF_PARALLEL) { tcg_gen_op1(ctx, INDEX_op_mb, mb_type); } } /* 32 bit ops */ void tcg_gen_movi_i32(TCGContext *s, TCGv_i32 ret, int32_t arg) { tcg_gen_mov_i32(s, ret, tcg_constant_i32(s, arg)); } void tcg_gen_addi_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_add_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_subfi_i32(TCGContext *s, TCGv_i32 ret, int32_t arg1, TCGv_i32 arg2) { if (arg1 == 0 && TCG_TARGET_HAS_neg_i32) { /* Don't recurse with tcg_gen_neg_i32. */ tcg_gen_op2_i32(s, INDEX_op_neg_i32, ret, arg2); } else { tcg_gen_sub_i32(s, ret, tcg_constant_i32(s, arg1), arg2); } } void tcg_gen_subi_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_sub_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_andi_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { /* Some cases can be optimized here. */ switch (arg2) { case 0: tcg_gen_movi_i32(s, ret, 0); return; case -1: tcg_gen_mov_i32(s, ret, arg1); return; case 0xff: /* Don't recurse with tcg_gen_ext8u_i32. */ if (TCG_TARGET_HAS_ext8u_i32) { tcg_gen_op2_i32(s, INDEX_op_ext8u_i32, ret, arg1); return; } break; case 0xffff: if (TCG_TARGET_HAS_ext16u_i32) { tcg_gen_op2_i32(s, INDEX_op_ext16u_i32, ret, arg1); return; } break; } tcg_gen_and_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } void tcg_gen_ori_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { /* Some cases can be optimized here. */ if (arg2 == -1) { tcg_gen_movi_i32(s, ret, -1); } else if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_or_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_xori_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { /* Some cases can be optimized here. */ if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else if (arg2 == -1 && TCG_TARGET_HAS_not_i32) { /* Don't recurse with tcg_gen_not_i32. */ tcg_gen_op2_i32(s, INDEX_op_not_i32, ret, arg1); } else { tcg_gen_xor_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_shli_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 32); if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_shl_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_shri_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 32); if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_shr_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_sari_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 32); if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_sar_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_brcond_i32(TCGContext *s, TCGCond cond, TCGv_i32 arg1, TCGv_i32 arg2, TCGLabel *l) { if (cond == TCG_COND_ALWAYS) { tcg_gen_br(s, l); } else if (cond != TCG_COND_NEVER) { l->refs++; tcg_gen_op4ii_i32(s, INDEX_op_brcond_i32, arg1, arg2, cond, label_arg(s, l)); } } void tcg_gen_brcondi_i32(TCGContext *s, TCGCond cond, TCGv_i32 arg1, int32_t arg2, TCGLabel *l) { if (cond == TCG_COND_ALWAYS) { tcg_gen_br(s, l); } else if (cond != TCG_COND_NEVER) { tcg_gen_brcond_i32(s, cond, arg1, tcg_constant_i32(s, arg2), l); } } void tcg_gen_setcond_i32(TCGContext *s, TCGCond cond, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (cond == TCG_COND_ALWAYS) { tcg_gen_movi_i32(s, ret, 1); } else if (cond == TCG_COND_NEVER) { tcg_gen_movi_i32(s, ret, 0); } else { tcg_gen_op4i_i32(s, INDEX_op_setcond_i32, ret, arg1, arg2, cond); } } void tcg_gen_setcondi_i32(TCGContext *s, TCGCond cond, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_gen_setcond_i32(s, cond, ret, arg1, tcg_constant_i32(s, arg2)); } void tcg_gen_muli_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { if (arg2 == 0) { tcg_gen_movi_i32(s, ret, 0); } else if (is_power_of_2(arg2)) { tcg_gen_shli_i32(s, ret, arg1, ctz32(arg2)); } else { tcg_gen_mul_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_div_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_div_i32) { tcg_gen_op3_i32(s, INDEX_op_div_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div2_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_sari_i32(s, t0, arg1, 31); tcg_gen_op5_i32(s, INDEX_op_div2_i32, ret, t0, arg1, t0, arg2); tcg_temp_free_i32(s, t0); } else { gen_helper_div_i32(s, ret, arg1, arg2); } } void tcg_gen_rem_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_rem_i32) { tcg_gen_op3_i32(s, INDEX_op_rem_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_op3_i32(s, INDEX_op_div_i32, t0, arg1, arg2); tcg_gen_mul_i32(s, t0, t0, arg2); tcg_gen_sub_i32(s, ret, arg1, t0); tcg_temp_free_i32(s, t0); } else if (TCG_TARGET_HAS_div2_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_sari_i32(s, t0, arg1, 31); tcg_gen_op5_i32(s, INDEX_op_div2_i32, t0, ret, arg1, t0, arg2); tcg_temp_free_i32(s, t0); } else { gen_helper_rem_i32(s, ret, arg1, arg2); } } void tcg_gen_divu_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_div_i32) { tcg_gen_op3_i32(s, INDEX_op_divu_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div2_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_movi_i32(s, t0, 0); tcg_gen_op5_i32(s, INDEX_op_divu2_i32, ret, t0, arg1, t0, arg2); tcg_temp_free_i32(s, t0); } else { gen_helper_divu_i32(s, ret, arg1, arg2); } } void tcg_gen_remu_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_rem_i32) { tcg_gen_op3_i32(s, INDEX_op_remu_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_op3_i32(s, INDEX_op_divu_i32, t0, arg1, arg2); tcg_gen_mul_i32(s, t0, t0, arg2); tcg_gen_sub_i32(s, ret, arg1, t0); tcg_temp_free_i32(s, t0); } else if (TCG_TARGET_HAS_div2_i32) { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_movi_i32(s, t0, 0); tcg_gen_op5_i32(s, INDEX_op_divu2_i32, t0, ret, arg1, t0, arg2); tcg_temp_free_i32(s, t0); } else { gen_helper_remu_i32(s, ret, arg1, arg2); } } void tcg_gen_andc_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_andc_i32) { tcg_gen_op3_i32(s, INDEX_op_andc_i32, ret, arg1, arg2); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_not_i32(s, t0, arg2); tcg_gen_and_i32(s, ret, arg1, t0); tcg_temp_free_i32(s, t0); } } void tcg_gen_eqv_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_eqv_i32) { tcg_gen_op3_i32(s, INDEX_op_eqv_i32, ret, arg1, arg2); } else { tcg_gen_xor_i32(s, ret, arg1, arg2); tcg_gen_not_i32(s, ret, ret); } } void tcg_gen_nand_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_nand_i32) { tcg_gen_op3_i32(s, INDEX_op_nand_i32, ret, arg1, arg2); } else { tcg_gen_and_i32(s, ret, arg1, arg2); tcg_gen_not_i32(s, ret, ret); } } void tcg_gen_nor_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_nor_i32) { tcg_gen_op3_i32(s, INDEX_op_nor_i32, ret, arg1, arg2); } else { tcg_gen_or_i32(s, ret, arg1, arg2); tcg_gen_not_i32(s, ret, ret); } } void tcg_gen_orc_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_orc_i32) { tcg_gen_op3_i32(s, INDEX_op_orc_i32, ret, arg1, arg2); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_not_i32(s, t0, arg2); tcg_gen_or_i32(s, ret, arg1, t0); tcg_temp_free_i32(s, t0); } } void tcg_gen_clz_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_clz_i32) { tcg_gen_op3_i32(s, INDEX_op_clz_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_clz_i64) { TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); tcg_gen_extu_i32_i64(s, t1, arg1); tcg_gen_extu_i32_i64(s, t2, arg2); tcg_gen_addi_i64(s, t2, t2, 32); tcg_gen_clz_i64(s, t1, t1, t2); tcg_gen_extrl_i64_i32(s, ret, t1); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); tcg_gen_subi_i32(s, ret, ret, 32); } else { gen_helper_clz_i32(s, ret, arg1, arg2); } } void tcg_gen_clzi_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, uint32_t arg2) { tcg_gen_clz_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } void tcg_gen_ctz_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_ctz_i32) { tcg_gen_op3_i32(s, INDEX_op_ctz_i32, ret, arg1, arg2); } else if (TCG_TARGET_HAS_ctz_i64) { TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); tcg_gen_extu_i32_i64(s, t1, arg1); tcg_gen_extu_i32_i64(s, t2, arg2); tcg_gen_ctz_i64(s, t1, t1, t2); tcg_gen_extrl_i64_i32(s, ret, t1); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); } else if (TCG_TARGET_HAS_ctpop_i32 || TCG_TARGET_HAS_ctpop_i64 || TCG_TARGET_HAS_clz_i32 || TCG_TARGET_HAS_clz_i64) { TCGv_i32 z, t = tcg_temp_new_i32(s); if (TCG_TARGET_HAS_ctpop_i32 || TCG_TARGET_HAS_ctpop_i64) { tcg_gen_subi_i32(s, t, arg1, 1); tcg_gen_andc_i32(s, t, t, arg1); tcg_gen_ctpop_i32(s, t, t); } else { /* Since all non-x86 hosts have clz(0) == 32, don't fight it. */ tcg_gen_neg_i32(s, t, arg1); tcg_gen_and_i32(s, t, t, arg1); tcg_gen_clzi_i32(s, t, t, 32); tcg_gen_xori_i32(s, t, t, 31); } z = tcg_constant_i32(s, 0); tcg_gen_movcond_i32(s, TCG_COND_EQ, ret, arg1, z, arg2, t); tcg_temp_free_i32(s, t); } else { gen_helper_ctz_i32(s, ret, arg1, arg2); } } void tcg_gen_ctzi_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, uint32_t arg2) { if (!TCG_TARGET_HAS_ctz_i32 && TCG_TARGET_HAS_ctpop_i32 && arg2 == 32) { /* This equivalence has the advantage of not requiring a fixup. */ TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_subi_i32(s, t, arg1, 1); tcg_gen_andc_i32(s, t, t, arg1); tcg_gen_ctpop_i32(s, ret, t); tcg_temp_free_i32(s, t); } else { tcg_gen_ctz_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } } void tcg_gen_clrsb_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_clz_i32) { TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_sari_i32(s, t, arg, 31); tcg_gen_xor_i32(s, t, t, arg); tcg_gen_clzi_i32(s, t, t, 32); tcg_gen_subi_i32(s, ret, t, 1); tcg_temp_free_i32(s, t); } else { gen_helper_clrsb_i32(s, ret, arg); } } void tcg_gen_ctpop_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1) { if (TCG_TARGET_HAS_ctpop_i32) { tcg_gen_op2_i32(s, INDEX_op_ctpop_i32, ret, arg1); } else if (TCG_TARGET_HAS_ctpop_i64) { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_extu_i32_i64(s, t, arg1); tcg_gen_ctpop_i64(s, t, t); tcg_gen_extrl_i64_i32(s, ret, t); tcg_temp_free_i64(s, t); } else { gen_helper_ctpop_i32(s, ret, arg1); } } void tcg_gen_rotl_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_rot_i32) { tcg_gen_op3_i32(s, INDEX_op_rotl_i32, ret, arg1, arg2); } else { TCGv_i32 t0, t1; t0 = tcg_temp_new_i32(s); t1 = tcg_temp_new_i32(s); tcg_gen_shl_i32(s, t0, arg1, arg2); tcg_gen_subfi_i32(s, t1, 32, arg2); tcg_gen_shr_i32(s, t1, arg1, t1); tcg_gen_or_i32(s, ret, t0, t1); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } } void tcg_gen_rotli_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 32); /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else if (TCG_TARGET_HAS_rot_i32) { tcg_gen_rotl_i32(s, ret, arg1, tcg_constant_i32(s, arg2)); } else { TCGv_i32 t0, t1; t0 = tcg_temp_new_i32(s); t1 = tcg_temp_new_i32(s); tcg_gen_shli_i32(s, t0, arg1, arg2); tcg_gen_shri_i32(s, t1, arg1, 32 - arg2); tcg_gen_or_i32(s, ret, t0, t1); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } } void tcg_gen_rotr_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_rot_i32) { tcg_gen_op3_i32(s, INDEX_op_rotr_i32, ret, arg1, arg2); } else { TCGv_i32 t0, t1; t0 = tcg_temp_new_i32(s); t1 = tcg_temp_new_i32(s); tcg_gen_shr_i32(s, t0, arg1, arg2); tcg_gen_subfi_i32(s, t1, 32, arg2); tcg_gen_shl_i32(s, t1, arg1, t1); tcg_gen_or_i32(s, ret, t0, t1); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } } void tcg_gen_rotri_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, int32_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 32); /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i32(s, ret, arg1); } else { tcg_gen_rotli_i32(s, ret, arg1, 32 - arg2); } } void tcg_gen_deposit_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg1, TCGv_i32 arg2, unsigned int ofs, unsigned int len) { uint32_t mask; TCGv_i32 t1; tcg_debug_assert(ofs < 32); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 32); tcg_debug_assert(ofs + len <= 32); if (len == 32) { tcg_gen_mov_i32(s, ret, arg2); return; } if (TCG_TARGET_HAS_deposit_i32 && TCG_TARGET_deposit_i32_valid(ofs, len)) { tcg_gen_op5ii_i32(s, INDEX_op_deposit_i32, ret, arg1, arg2, ofs, len); return; } t1 = tcg_temp_new_i32(s); if (TCG_TARGET_HAS_extract2_i32) { if (ofs + len == 32) { tcg_gen_shli_i32(s, t1, arg1, len); tcg_gen_extract2_i32(s, ret, t1, arg2, len); goto done; } if (ofs == 0) { tcg_gen_extract2_i32(s, ret, arg1, arg2, len); tcg_gen_rotli_i32(s, ret, ret, len); goto done; } } mask = (1u << len) - 1; if (ofs + len < 32) { tcg_gen_andi_i32(s, t1, arg2, mask); tcg_gen_shli_i32(s, t1, t1, ofs); } else { tcg_gen_shli_i32(s, t1, arg2, ofs); } tcg_gen_andi_i32(s, ret, arg1, ~(mask << ofs)); tcg_gen_or_i32(s, ret, ret, t1); done: tcg_temp_free_i32(s, t1); } void tcg_gen_deposit_z_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 32); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 32); tcg_debug_assert(ofs + len <= 32); if (ofs + len == 32) { tcg_gen_shli_i32(s, ret, arg, ofs); } else if (ofs == 0) { tcg_gen_andi_i32(s, ret, arg, (1u << len) - 1); } else if (TCG_TARGET_HAS_deposit_i32 && TCG_TARGET_deposit_i32_valid(ofs, len)) { TCGv_i32 zero = tcg_constant_i32(s, 0); tcg_gen_op5ii_i32(s, INDEX_op_deposit_i32, ret, zero, arg, ofs, len); } else { /* To help two-operand hosts we prefer to zero-extend first, which allows ARG to stay live. */ switch (len) { case 16: if (TCG_TARGET_HAS_ext16u_i32) { tcg_gen_ext16u_i32(s, ret, arg); tcg_gen_shli_i32(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i32) { tcg_gen_ext8u_i32(s, ret, arg); tcg_gen_shli_i32(s, ret, ret, ofs); return; } break; } /* Otherwise prefer zero-extension over AND for code size. */ switch (ofs + len) { case 16: if (TCG_TARGET_HAS_ext16u_i32) { tcg_gen_shli_i32(s, ret, arg, ofs); tcg_gen_ext16u_i32(s, ret, ret); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i32) { tcg_gen_shli_i32(s, ret, arg, ofs); tcg_gen_ext8u_i32(s, ret, ret); return; } break; } tcg_gen_andi_i32(s, ret, arg, (1u << len) - 1); tcg_gen_shli_i32(s, ret, ret, ofs); } } void tcg_gen_extract_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 32); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 32); tcg_debug_assert(ofs + len <= 32); /* Canonicalize certain special cases, even if extract is supported. */ if (ofs + len == 32) { tcg_gen_shri_i32(s, ret, arg, 32 - len); return; } if (ofs == 0) { tcg_gen_andi_i32(s, ret, arg, (1u << len) - 1); return; } if (TCG_TARGET_HAS_extract_i32 && TCG_TARGET_extract_i32_valid(ofs, len)) { tcg_gen_op4ii_i32(s, INDEX_op_extract_i32, ret, arg, ofs, len); return; } /* Assume that zero-extension, if available, is cheaper than a shift. */ switch (ofs + len) { case 16: if (TCG_TARGET_HAS_ext16u_i32) { tcg_gen_ext16u_i32(s, ret, arg); tcg_gen_shri_i32(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i32) { tcg_gen_ext8u_i32(s, ret, arg); tcg_gen_shri_i32(s, ret, ret, ofs); return; } break; } /* ??? Ideally we'd know what values are available for immediate AND. Assume that 8 bits are available, plus the special case of 16, so that we get ext8u, ext16u. */ switch (len) { case 1: case 2: case 3: case 4: case 5: case 6: case 7: case 8: case 16: tcg_gen_shri_i32(s, ret, arg, ofs); tcg_gen_andi_i32(s, ret, ret, (1u << len) - 1); break; default: tcg_gen_shli_i32(s, ret, arg, 32 - len - ofs); tcg_gen_shri_i32(s, ret, ret, 32 - len); break; } } void tcg_gen_sextract_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 32); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 32); tcg_debug_assert(ofs + len <= 32); /* Canonicalize certain special cases, even if extract is supported. */ if (ofs + len == 32) { tcg_gen_sari_i32(s, ret, arg, 32 - len); return; } if (ofs == 0) { switch (len) { case 16: tcg_gen_ext16s_i32(s, ret, arg); return; case 8: tcg_gen_ext8s_i32(s, ret, arg); return; } } if (TCG_TARGET_HAS_sextract_i32 && TCG_TARGET_extract_i32_valid(ofs, len)) { tcg_gen_op4ii_i32(s, INDEX_op_sextract_i32, ret, arg, ofs, len); return; } /* Assume that sign-extension, if available, is cheaper than a shift. */ switch (ofs + len) { case 16: if (TCG_TARGET_HAS_ext16s_i32) { tcg_gen_ext16s_i32(s, ret, arg); tcg_gen_sari_i32(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8s_i32) { tcg_gen_ext8s_i32(s, ret, arg); tcg_gen_sari_i32(s, ret, ret, ofs); return; } break; } switch (len) { case 16: if (TCG_TARGET_HAS_ext16s_i32) { tcg_gen_shri_i32(s, ret, arg, ofs); tcg_gen_ext16s_i32(s, ret, ret); return; } break; case 8: if (TCG_TARGET_HAS_ext8s_i32) { tcg_gen_shri_i32(s, ret, arg, ofs); tcg_gen_ext8s_i32(s, ret, ret); return; } break; } tcg_gen_shli_i32(s, ret, arg, 32 - len - ofs); tcg_gen_sari_i32(s, ret, ret, 32 - len); } /* * Extract 32-bits from a 64-bit input, ah:al, starting from ofs. * Unlike tcg_gen_extract_i32 above, len is fixed at 32. */ void tcg_gen_extract2_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 al, TCGv_i32 ah, unsigned int ofs) { tcg_debug_assert(ofs <= 32); if (ofs == 0) { tcg_gen_mov_i32(s, ret, al); } else if (ofs == 32) { tcg_gen_mov_i32(s, ret, ah); } else if (al == ah) { tcg_gen_rotri_i32(s, ret, al, ofs); } else if (TCG_TARGET_HAS_extract2_i32) { tcg_gen_op4i_i32(s, INDEX_op_extract2_i32, ret, al, ah, ofs); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_shri_i32(s, t0, al, ofs); tcg_gen_deposit_i32(s, ret, t0, ah, 32 - ofs, ofs); tcg_temp_free_i32(s, t0); } } void tcg_gen_movcond_i32(TCGContext *s, TCGCond cond, TCGv_i32 ret, TCGv_i32 c1, TCGv_i32 c2, TCGv_i32 v1, TCGv_i32 v2) { if (cond == TCG_COND_ALWAYS) { tcg_gen_mov_i32(s, ret, v1); } else if (cond == TCG_COND_NEVER) { tcg_gen_mov_i32(s, ret, v2); } else if (TCG_TARGET_HAS_movcond_i32) { tcg_gen_op6i_i32(s, INDEX_op_movcond_i32, ret, c1, c2, v1, v2, cond); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); TCGv_i32 t1 = tcg_temp_new_i32(s); tcg_gen_setcond_i32(s, cond, t0, c1, c2); tcg_gen_neg_i32(s, t0, t0); tcg_gen_and_i32(s, t1, v1, t0); tcg_gen_andc_i32(s, ret, v2, t0); tcg_gen_or_i32(s, ret, ret, t1); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } } void tcg_gen_add2_i32(TCGContext *s, TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 al, TCGv_i32 ah, TCGv_i32 bl, TCGv_i32 bh) { if (TCG_TARGET_HAS_add2_i32) { tcg_gen_op6_i32(s, INDEX_op_add2_i32, rl, rh, al, ah, bl, bh); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_concat_i32_i64(s, t0, al, ah); tcg_gen_concat_i32_i64(s, t1, bl, bh); tcg_gen_add_i64(s, t0, t0, t1); tcg_gen_extr_i64_i32(s, rl, rh, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_sub2_i32(TCGContext *s, TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 al, TCGv_i32 ah, TCGv_i32 bl, TCGv_i32 bh) { if (TCG_TARGET_HAS_sub2_i32) { tcg_gen_op6_i32(s, INDEX_op_sub2_i32, rl, rh, al, ah, bl, bh); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_concat_i32_i64(s, t0, al, ah); tcg_gen_concat_i32_i64(s, t1, bl, bh); tcg_gen_sub_i64(s, t0, t0, t1); tcg_gen_extr_i64_i32(s, rl, rh, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_mulu2_i32(TCGContext *s, TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_mulu2_i32) { tcg_gen_op4_i32(s, INDEX_op_mulu2_i32, rl, rh, arg1, arg2); } else if (TCG_TARGET_HAS_muluh_i32) { TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_op3_i32(s, INDEX_op_mul_i32, t, arg1, arg2); tcg_gen_op3_i32(s, INDEX_op_muluh_i32, rh, arg1, arg2); tcg_gen_mov_i32(s, rl, t); tcg_temp_free_i32(s, t); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_extu_i32_i64(s, t0, arg1); tcg_gen_extu_i32_i64(s, t1, arg2); tcg_gen_mul_i64(s, t0, t0, t1); tcg_gen_extr_i64_i32(s, rl, rh, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_muls2_i32(TCGContext *s, TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_HAS_muls2_i32) { tcg_gen_op4_i32(s, INDEX_op_muls2_i32, rl, rh, arg1, arg2); } else if (TCG_TARGET_HAS_mulsh_i32) { TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_op3_i32(s, INDEX_op_mul_i32, t, arg1, arg2); tcg_gen_op3_i32(s, INDEX_op_mulsh_i32, rh, arg1, arg2); tcg_gen_mov_i32(s, rl, t); tcg_temp_free_i32(s, t); } else if (TCG_TARGET_REG_BITS == 32) { TCGv_i32 t0 = tcg_temp_new_i32(s); TCGv_i32 t1 = tcg_temp_new_i32(s); TCGv_i32 t2 = tcg_temp_new_i32(s); TCGv_i32 t3 = tcg_temp_new_i32(s); tcg_gen_mulu2_i32(s, t0, t1, arg1, arg2); /* Adjust for negative inputs. */ tcg_gen_sari_i32(s, t2, arg1, 31); tcg_gen_sari_i32(s, t3, arg2, 31); tcg_gen_and_i32(s, t2, t2, arg2); tcg_gen_and_i32(s, t3, t3, arg1); tcg_gen_sub_i32(s, rh, t1, t2); tcg_gen_sub_i32(s, rh, rh, t3); tcg_gen_mov_i32(s, rl, t0); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); tcg_temp_free_i32(s, t2); tcg_temp_free_i32(s, t3); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_ext_i32_i64(s, t0, arg1); tcg_gen_ext_i32_i64(s, t1, arg2); tcg_gen_mul_i64(s, t0, t0, t1); tcg_gen_extr_i64_i32(s, rl, rh, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_mulsu2_i32(TCGContext *s, TCGv_i32 rl, TCGv_i32 rh, TCGv_i32 arg1, TCGv_i32 arg2) { if (TCG_TARGET_REG_BITS == 32) { TCGv_i32 t0 = tcg_temp_new_i32(s); TCGv_i32 t1 = tcg_temp_new_i32(s); TCGv_i32 t2 = tcg_temp_new_i32(s); tcg_gen_mulu2_i32(s, t0, t1, arg1, arg2); /* Adjust for negative input for the signed arg1. */ tcg_gen_sari_i32(s, t2, arg1, 31); tcg_gen_and_i32(s, t2, t2, arg2); tcg_gen_sub_i32(s, rh, t1, t2); tcg_gen_mov_i32(s, rl, t0); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); tcg_temp_free_i32(s, t2); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_ext_i32_i64(s, t0, arg1); tcg_gen_extu_i32_i64(s, t1, arg2); tcg_gen_mul_i64(s, t0, t0, t1); tcg_gen_extr_i64_i32(s, rl, rh, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_ext8s_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_ext8s_i32) { tcg_gen_op2_i32(s, INDEX_op_ext8s_i32, ret, arg); } else { tcg_gen_shli_i32(s, ret, arg, 24); tcg_gen_sari_i32(s, ret, ret, 24); } } void tcg_gen_ext16s_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_ext16s_i32) { tcg_gen_op2_i32(s, INDEX_op_ext16s_i32, ret, arg); } else { tcg_gen_shli_i32(s, ret, arg, 16); tcg_gen_sari_i32(s, ret, ret, 16); } } void tcg_gen_ext8u_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_ext8u_i32) { tcg_gen_op2_i32(s, INDEX_op_ext8u_i32, ret, arg); } else { tcg_gen_andi_i32(s, ret, arg, 0xffu); } } void tcg_gen_ext16u_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_ext16u_i32) { tcg_gen_op2_i32(s, INDEX_op_ext16u_i32, ret, arg); } else { tcg_gen_andi_i32(s, ret, arg, 0xffffu); } } /* Note: we assume the two high bytes are set to zero */ void tcg_gen_bswap16_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_bswap16_i32) { tcg_gen_op2_i32(s, INDEX_op_bswap16_i32, ret, arg); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_ext8u_i32(s, t0, arg); tcg_gen_shli_i32(s, t0, t0, 8); tcg_gen_shri_i32(s, ret, arg, 8); tcg_gen_or_i32(s, ret, ret, t0); tcg_temp_free_i32(s, t0); } } void tcg_gen_bswap32_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 arg) { if (TCG_TARGET_HAS_bswap32_i32) { tcg_gen_op2_i32(s, INDEX_op_bswap32_i32, ret, arg); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); TCGv_i32 t1 = tcg_temp_new_i32(s); TCGv_i32 t2 = tcg_constant_i32(s, 0x00ff00ff); /* arg = abcd */ tcg_gen_shri_i32(s, t0, arg, 8); /* t0 = .abc */ tcg_gen_and_i32(s, t1, arg, t2); /* t1 = .b.d */ tcg_gen_and_i32(s, t0, t0, t2); /* t0 = .a.c */ tcg_gen_shli_i32(s, t1, t1, 8); /* t1 = b.d. */ tcg_gen_or_i32(s, ret, t0, t1); /* ret = badc */ tcg_gen_shri_i32(s, t0, ret, 16); /* t0 = ..ba */ tcg_gen_shli_i32(s, t1, ret, 16); /* t1 = dc.. */ tcg_gen_or_i32(s, ret, t0, t1); /* ret = dcba */ tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } } void tcg_gen_smin_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 a, TCGv_i32 b) { tcg_gen_movcond_i32(s, TCG_COND_LT, ret, a, b, a, b); } void tcg_gen_umin_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 a, TCGv_i32 b) { tcg_gen_movcond_i32(s, TCG_COND_LTU, ret, a, b, a, b); } void tcg_gen_smax_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 a, TCGv_i32 b) { tcg_gen_movcond_i32(s, TCG_COND_LT, ret, a, b, b, a); } void tcg_gen_umax_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 a, TCGv_i32 b) { tcg_gen_movcond_i32(s, TCG_COND_LTU, ret, a, b, b, a); } void tcg_gen_abs_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 a) { TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_sari_i32(s, t, a, 31); tcg_gen_xor_i32(s, ret, a, t); tcg_gen_sub_i32(s, ret, ret, t); tcg_temp_free_i32(s, t); } /* 64-bit ops */ #if TCG_TARGET_REG_BITS == 32 /* These are all inline for TCG_TARGET_REG_BITS == 64. */ void tcg_gen_discard_i64(TCGContext *s, TCGv_i64 arg) { tcg_gen_discard_i32(s, TCGV_LOW(s, arg)); tcg_gen_discard_i32(s, TCGV_HIGH(s, arg)); } void tcg_gen_mov_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { TCGTemp *ts = tcgv_i64_temp(s, arg); /* Canonicalize TCGv_i64 TEMP_CONST into TCGv_i32 TEMP_CONST. */ if (ts->kind == TEMP_CONST) { tcg_gen_movi_i64(s, ret, ts->val); } else { tcg_gen_mov_i32(s, TCGV_LOW(ret), TCGV_LOW(arg)); tcg_gen_mov_i32(s, TCGV_HIGH(ret), TCGV_HIGH(arg)); } } void tcg_gen_movi_i64(TCGContext *s, TCGv_i64 ret, int64_t arg) { tcg_gen_movi_i32(s, TCGV_LOW(s, ret), arg); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), arg >> 32); } void tcg_gen_ld8u_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld8u_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } void tcg_gen_ld8s_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld8s_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } void tcg_gen_ld16u_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld16u_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } void tcg_gen_ld16s_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld16s_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } void tcg_gen_ld32u_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } void tcg_gen_ld32s_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { tcg_gen_ld_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } void tcg_gen_ld_i64(TCGContext *s, TCGv_i64 ret, TCGv_ptr arg2, tcg_target_long offset) { /* Since arg2 and ret have different types, they cannot be the same temporary */ #ifdef HOST_WORDS_BIGENDIAN tcg_gen_ld_i32(s, TCGV_HIGH(s, ret), arg2, offset); tcg_gen_ld_i32(s, TCGV_LOW(s, ret), arg2, offset + 4); #else tcg_gen_ld_i32(s, TCGV_LOW(s, ret), arg2, offset); tcg_gen_ld_i32(s, TCGV_HIGH(s, ret), arg2, offset + 4); #endif } void tcg_gen_st_i64(TCGContext *s, TCGv_i64 arg1, TCGv_ptr arg2, tcg_target_long offset) { #ifdef HOST_WORDS_BIGENDIAN tcg_gen_st_i32(s, TCGV_HIGH(s, arg1), arg2, offset); tcg_gen_st_i32(s, TCGV_LOW(s, arg1), arg2, offset + 4); #else tcg_gen_st_i32(s, TCGV_LOW(s, arg1), arg2, offset); tcg_gen_st_i32(s, TCGV_HIGH(s, arg1), arg2, offset + 4); #endif } void tcg_gen_and_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { tcg_gen_and_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_and_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } void tcg_gen_or_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { tcg_gen_or_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_or_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } void tcg_gen_xor_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { tcg_gen_xor_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_xor_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } void tcg_gen_shl_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { gen_helper_shl_i64(s, ret, arg1, arg2); } void tcg_gen_shr_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { gen_helper_shr_i64(s, ret, arg1, arg2); } void tcg_gen_sar_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { gen_helper_sar_i64(s, ret, arg1, arg2); } void tcg_gen_mul_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { TCGv_i64 t0; TCGv_i32 t1; t0 = tcg_temp_new_i64(s); t1 = tcg_temp_new_i32(s); tcg_gen_mulu2_i32(s, TCGV_LOW(s, t0), TCGV_HIGH(s, t0), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_mul_i32(s, t1, TCGV_LOW(s, arg1), TCGV_HIGH(s, arg2)); tcg_gen_add_i32(s, TCGV_HIGH(s, t0), TCGV_HIGH(s, t0), t1); tcg_gen_mul_i32(s, t1, TCGV_HIGH(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_add_i32(s, TCGV_HIGH(s, t0), TCGV_HIGH(s, t0), t1); tcg_gen_mov_i64(s, ret, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i32(s, t1); } #else void tcg_gen_movi_i64(TCGContext *s, TCGv_i64 ret, int64_t arg) { tcg_gen_mov_i64(s, ret, tcg_constant_i64(s, arg)); } #endif /* TCG_TARGET_REG_SIZE == 32 */ void tcg_gen_addi_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else if (TCG_TARGET_REG_BITS == 64) { tcg_gen_add_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } else { tcg_gen_add2_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), tcg_constant_i32(s, arg2), tcg_constant_i32(s, arg2 >> 32)); } } void tcg_gen_subfi_i64(TCGContext *s, TCGv_i64 ret, int64_t arg1, TCGv_i64 arg2) { if (arg1 == 0 && TCG_TARGET_HAS_neg_i64) { /* Don't recurse with tcg_gen_neg_i64. */ tcg_gen_op2_i64(s, INDEX_op_neg_i64, ret, arg2); } else if (TCG_TARGET_REG_BITS == 64) { tcg_gen_sub_i64(s, ret, tcg_constant_i64(s, arg1), arg2); } else { tcg_gen_sub2_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, ret), tcg_constant_i32(s, arg1), tcg_constant_i32(s, arg1 >> 32), TCGV_LOW(s, arg2), TCGV_HIGH(s, arg2)); } } void tcg_gen_subi_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else if (TCG_TARGET_REG_BITS == 64) { tcg_gen_sub_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } else { tcg_gen_sub2_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), tcg_constant_i32(s, arg2), tcg_constant_i32(s, arg2 >> 32)); } } void tcg_gen_andi_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_andi_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), arg2); tcg_gen_andi_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), arg2 >> 32); return; } /* Some cases can be optimized here. */ switch (arg2) { case 0: tcg_gen_movi_i64(s, ret, 0); return; case -1: tcg_gen_mov_i64(s, ret, arg1); return; case 0xff: /* Don't recurse with tcg_gen_ext8u_i64. */ if (TCG_TARGET_HAS_ext8u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext8u_i64, ret, arg1); return; } break; case 0xffff: if (TCG_TARGET_HAS_ext16u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext16u_i64, ret, arg1); return; } break; case 0xffffffff: if (TCG_TARGET_HAS_ext32u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext32u_i64, ret, arg1); return; } break; } tcg_gen_and_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } void tcg_gen_ori_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_ori_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), arg2); tcg_gen_ori_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), arg2 >> 32); return; } /* Some cases can be optimized here. */ if (arg2 == -1) { tcg_gen_movi_i64(s, ret, -1); } else if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else { tcg_gen_or_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } } void tcg_gen_xori_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_xori_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), arg2); tcg_gen_xori_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), arg2 >> 32); return; } /* Some cases can be optimized here. */ if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else if (arg2 == -1 && TCG_TARGET_HAS_not_i64) { /* Don't recurse with tcg_gen_not_i64. */ tcg_gen_op2_i64(s, INDEX_op_not_i64, ret, arg1); } else { tcg_gen_xor_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } } static inline void tcg_gen_shifti_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, unsigned c, bool right, bool arith) { tcg_debug_assert(c < 64); if (c == 0) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1)); tcg_gen_mov_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1)); } else if (c >= 32) { c -= 32; if (right) { if (arith) { tcg_gen_sari_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, arg1), c); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), 31); } else { tcg_gen_shri_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, arg1), c); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } } else { tcg_gen_shli_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, arg1), c); tcg_gen_movi_i32(s, TCGV_LOW(s, ret), 0); } } else if (right) { if (TCG_TARGET_HAS_extract2_i32) { tcg_gen_extract2_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), c); } else { tcg_gen_shri_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), c); tcg_gen_deposit_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, ret), TCGV_HIGH(s, arg1), 32 - c, c); } if (arith) { tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), c); } else { tcg_gen_shri_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), c); } } else { if (TCG_TARGET_HAS_extract2_i32) { tcg_gen_extract2_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), 32 - c); } else { TCGv_i32 t0 = tcg_temp_new_i32(s); tcg_gen_shri_i32(s, t0, TCGV_LOW(s, arg1), 32 - c); tcg_gen_deposit_i32(s, TCGV_HIGH(s, ret), t0, TCGV_HIGH(s, arg1), c, 32 - c); tcg_temp_free_i32(s, t0); } tcg_gen_shli_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), c); } } void tcg_gen_shli_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 64); if (TCG_TARGET_REG_BITS == 32) { tcg_gen_shifti_i64(s, ret, arg1, arg2, 0, 0); } else if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else { tcg_gen_shl_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } } void tcg_gen_shri_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 64); if (TCG_TARGET_REG_BITS == 32) { tcg_gen_shifti_i64(s, ret, arg1, arg2, 1, 0); } else if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else { tcg_gen_shr_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } } void tcg_gen_sari_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 64); if (TCG_TARGET_REG_BITS == 32) { tcg_gen_shifti_i64(s, ret, arg1, arg2, 1, 1); } else if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else { tcg_gen_sar_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } } void tcg_gen_brcond_i64(TCGContext *s, TCGCond cond, TCGv_i64 arg1, TCGv_i64 arg2, TCGLabel *l) { if (cond == TCG_COND_ALWAYS) { tcg_gen_br(s, l); } else if (cond != TCG_COND_NEVER) { l->refs++; if (TCG_TARGET_REG_BITS == 32) { tcg_gen_op6ii_i32(s, INDEX_op_brcond2_i32, TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), TCGV_LOW(s, arg2), TCGV_HIGH(s, arg2), cond, label_arg(s, l)); } else { tcg_gen_op4ii_i64(s, INDEX_op_brcond_i64, arg1, arg2, cond, label_arg(s, l)); } } } void tcg_gen_brcondi_i64(TCGContext *s, TCGCond cond, TCGv_i64 arg1, int64_t arg2, TCGLabel *l) { if (TCG_TARGET_REG_BITS == 64) { tcg_gen_brcond_i64(s, cond, arg1, tcg_constant_i64(s, arg2), l); } else if (cond == TCG_COND_ALWAYS) { tcg_gen_br(s, l); } else if (cond != TCG_COND_NEVER) { l->refs++; tcg_gen_op6ii_i32(s, INDEX_op_brcond2_i32, TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), tcg_constant_i32(s, arg2), tcg_constant_i32(s, arg2 >> 32), cond, label_arg(s, l)); } } void tcg_gen_setcond_i64(TCGContext *s, TCGCond cond, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (cond == TCG_COND_ALWAYS) { tcg_gen_movi_i64(s, ret, 1); } else if (cond == TCG_COND_NEVER) { tcg_gen_movi_i64(s, ret, 0); } else { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_op6i_i32(s, INDEX_op_setcond2_i32, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), TCGV_LOW(s, arg2), TCGV_HIGH(s, arg2), cond); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else { tcg_gen_op4i_i64(s, INDEX_op_setcond_i64, ret, arg1, arg2, cond); } } } void tcg_gen_setcondi_i64(TCGContext *s, TCGCond cond, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { if (TCG_TARGET_REG_BITS == 64) { tcg_gen_setcond_i64(s, cond, ret, arg1, tcg_constant_i64(s, arg2)); } else if (cond == TCG_COND_ALWAYS) { tcg_gen_movi_i64(s, ret, 1); } else if (cond == TCG_COND_NEVER) { tcg_gen_movi_i64(s, ret, 0); } else { tcg_gen_op6i_i32(s, INDEX_op_setcond2_i32, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_HIGH(s, arg1), tcg_constant_i32(s, arg2), tcg_constant_i32(s, arg2 >> 32), cond); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } } void tcg_gen_muli_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { if (arg2 == 0) { tcg_gen_movi_i64(s, ret, 0); } else if (is_power_of_2(arg2)) { tcg_gen_shli_i64(s, ret, arg1, ctz64(arg2)); } else { TCGv_i64 t0 = tcg_const_i64(s, arg2); tcg_gen_mul_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_div_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_div_i64) { tcg_gen_op3_i64(s, INDEX_op_div_i64, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div2_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_sari_i64(s, t0, arg1, 63); tcg_gen_op5_i64(s, INDEX_op_div2_i64, ret, t0, arg1, t0, arg2); tcg_temp_free_i64(s, t0); } else { gen_helper_div_i64(s, ret, arg1, arg2); } } void tcg_gen_rem_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_rem_i64) { tcg_gen_op3_i64(s, INDEX_op_rem_i64, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_op3_i64(s, INDEX_op_div_i64, t0, arg1, arg2); tcg_gen_mul_i64(s, t0, t0, arg2); tcg_gen_sub_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } else if (TCG_TARGET_HAS_div2_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_sari_i64(s, t0, arg1, 63); tcg_gen_op5_i64(s, INDEX_op_div2_i64, t0, ret, arg1, t0, arg2); tcg_temp_free_i64(s, t0); } else { gen_helper_rem_i64(s, ret, arg1, arg2); } } void tcg_gen_divu_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_div_i64) { tcg_gen_op3_i64(s, INDEX_op_divu_i64, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div2_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_movi_i64(s, t0, 0); tcg_gen_op5_i64(s, INDEX_op_divu2_i64, ret, t0, arg1, t0, arg2); tcg_temp_free_i64(s, t0); } else { gen_helper_divu_i64(s, ret, arg1, arg2); } } void tcg_gen_remu_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_rem_i64) { tcg_gen_op3_i64(s, INDEX_op_remu_i64, ret, arg1, arg2); } else if (TCG_TARGET_HAS_div_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_op3_i64(s, INDEX_op_divu_i64, t0, arg1, arg2); tcg_gen_mul_i64(s, t0, t0, arg2); tcg_gen_sub_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } else if (TCG_TARGET_HAS_div2_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_movi_i64(s, t0, 0); tcg_gen_op5_i64(s, INDEX_op_divu2_i64, t0, ret, arg1, t0, arg2); tcg_temp_free_i64(s, t0); } else { gen_helper_remu_i64(s, ret, arg1, arg2); } } void tcg_gen_ext8s_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_ext8s_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } else if (TCG_TARGET_HAS_ext8s_i64) { tcg_gen_op2_i64(s, INDEX_op_ext8s_i64, ret, arg); } else { tcg_gen_shli_i64(s, ret, arg, 56); tcg_gen_sari_i64(s, ret, ret, 56); } } void tcg_gen_ext16s_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_ext16s_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } else if (TCG_TARGET_HAS_ext16s_i64) { tcg_gen_op2_i64(s, INDEX_op_ext16s_i64, ret, arg); } else { tcg_gen_shli_i64(s, ret, arg, 48); tcg_gen_sari_i64(s, ret, ret, 48); } } void tcg_gen_ext32s_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } else if (TCG_TARGET_HAS_ext32s_i64) { tcg_gen_op2_i64(s, INDEX_op_ext32s_i64, ret, arg); } else { tcg_gen_shli_i64(s, ret, arg, 32); tcg_gen_sari_i64(s, ret, ret, 32); } } void tcg_gen_ext8u_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_ext8u_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else if (TCG_TARGET_HAS_ext8u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext8u_i64, ret, arg); } else { tcg_gen_andi_i64(s, ret, arg, 0xffu); } } void tcg_gen_ext16u_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_ext16u_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else if (TCG_TARGET_HAS_ext16u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext16u_i64, ret, arg); } else { tcg_gen_andi_i64(s, ret, arg, 0xffffu); } } void tcg_gen_ext32u_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else if (TCG_TARGET_HAS_ext32u_i64) { tcg_gen_op2_i64(s, INDEX_op_ext32u_i64, ret, arg); } else { tcg_gen_andi_i64(s, ret, arg, 0xffffffffu); } } /* Note: we assume the six high bytes are set to zero */ void tcg_gen_bswap16_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_bswap16_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else if (TCG_TARGET_HAS_bswap16_i64) { tcg_gen_op2_i64(s, INDEX_op_bswap16_i64, ret, arg); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_ext8u_i64(s, t0, arg); tcg_gen_shli_i64(s, t0, t0, 8); tcg_gen_shri_i64(s, ret, arg, 8); tcg_gen_or_i64(s, ret, ret, t0); tcg_temp_free_i64(s, t0); } } /* Note: we assume the four high bytes are set to zero */ void tcg_gen_bswap32_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_bswap32_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else if (TCG_TARGET_HAS_bswap32_i64) { tcg_gen_op2_i64(s, INDEX_op_bswap32_i64, ret, arg); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_constant_i64(s, 0x00ff00ff); /* arg = ....abcd */ tcg_gen_shri_i64(s, t0, arg, 8); /* t0 = .....abc */ tcg_gen_and_i64(s, t1, arg, t2); /* t1 = .....b.d */ tcg_gen_and_i64(s, t0, t0, t2); /* t0 = .....a.c */ tcg_gen_shli_i64(s, t1, t1, 8); /* t1 = ....b.d. */ tcg_gen_or_i64(s, ret, t0, t1); /* ret = ....badc */ tcg_gen_shli_i64(s, t1, ret, 48); /* t1 = dc...... */ tcg_gen_shri_i64(s, t0, ret, 16); /* t0 = ......ba */ tcg_gen_shri_i64(s, t1, t1, 32); /* t1 = ....dc.. */ tcg_gen_or_i64(s, ret, t0, t1); /* ret = ....dcba */ tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_bswap64_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { TCGv_i32 t0, t1; t0 = tcg_temp_new_i32(s); t1 = tcg_temp_new_i32(s); tcg_gen_bswap32_i32(s, t0, TCGV_LOW(s, arg)); tcg_gen_bswap32_i32(s, t1, TCGV_HIGH(s, arg)); tcg_gen_mov_i32(s, TCGV_LOW(s, ret), t1); tcg_gen_mov_i32(s, TCGV_HIGH(s, ret), t0); tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } else if (TCG_TARGET_HAS_bswap64_i64) { tcg_gen_op2_i64(s, INDEX_op_bswap64_i64, ret, arg); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); /* arg = abcdefgh */ tcg_gen_movi_i64(s, t2, 0x00ff00ff00ff00ffull); tcg_gen_shri_i64(s, t0, arg, 8); /* t0 = .abcdefg */ tcg_gen_and_i64(s, t1, arg, t2); /* t1 = .b.d.f.h */ tcg_gen_and_i64(s, t0, t0, t2); /* t0 = .a.c.e.g */ tcg_gen_shli_i64(s, t1, t1, 8); /* t1 = b.d.f.h. */ tcg_gen_or_i64(s, ret, t0, t1); /* ret = badcfehg */ tcg_gen_movi_i64(s, t2, 0x0000ffff0000ffffull); tcg_gen_shri_i64(s, t0, ret, 16); /* t0 = ..badcfe */ tcg_gen_and_i64(s, t1, ret, t2); /* t1 = ..dc..hg */ tcg_gen_and_i64(s, t0, t0, t2); /* t0 = ..ba..fe */ tcg_gen_shli_i64(s, t1, t1, 16); /* t1 = dc..hg.. */ tcg_gen_or_i64(s, ret, t0, t1); /* ret = dcbahgfe */ tcg_gen_shri_i64(s, t0, ret, 32); /* t0 = ....dcba */ tcg_gen_shli_i64(s, t1, ret, 32); /* t1 = hgfe.... */ tcg_gen_or_i64(s, ret, t0, t1); /* ret = hgfedcba */ tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); } } void tcg_gen_not_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_not_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_not_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg)); } else if (TCG_TARGET_HAS_not_i64) { tcg_gen_op2_i64(s, INDEX_op_not_i64, ret, arg); } else { tcg_gen_xori_i64(s, ret, arg, -1); } } void tcg_gen_andc_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_andc_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_andc_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } else if (TCG_TARGET_HAS_andc_i64) { tcg_gen_op3_i64(s, INDEX_op_andc_i64, ret, arg1, arg2); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_not_i64(s, t0, arg2); tcg_gen_and_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_eqv_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_eqv_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_eqv_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } else if (TCG_TARGET_HAS_eqv_i64) { tcg_gen_op3_i64(s, INDEX_op_eqv_i64, ret, arg1, arg2); } else { tcg_gen_xor_i64(s, ret, arg1, arg2); tcg_gen_not_i64(s, ret, ret); } } void tcg_gen_nand_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_nand_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_nand_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } else if (TCG_TARGET_HAS_nand_i64) { tcg_gen_op3_i64(s, INDEX_op_nand_i64, ret, arg1, arg2); } else { tcg_gen_and_i64(s, ret, arg1, arg2); tcg_gen_not_i64(s, ret, ret); } } void tcg_gen_nor_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_nor_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_nor_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } else if (TCG_TARGET_HAS_nor_i64) { tcg_gen_op3_i64(s, INDEX_op_nor_i64, ret, arg1, arg2); } else { tcg_gen_or_i64(s, ret, arg1, arg2); tcg_gen_not_i64(s, ret, ret); } } void tcg_gen_orc_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_orc_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2)); tcg_gen_orc_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_HIGH(s, arg2)); } else if (TCG_TARGET_HAS_orc_i64) { tcg_gen_op3_i64(s, INDEX_op_orc_i64, ret, arg1, arg2); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_not_i64(s, t0, arg2); tcg_gen_or_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_clz_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_clz_i64) { tcg_gen_op3_i64(s, INDEX_op_clz_i64, ret, arg1, arg2); } else { gen_helper_clz_i64(s, ret, arg1, arg2); } } void tcg_gen_clzi_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, uint64_t arg2) { if (TCG_TARGET_REG_BITS == 32 && TCG_TARGET_HAS_clz_i32 && arg2 <= 0xffffffffu) { TCGv_i32 t = tcg_temp_new_i32(s); tcg_gen_clzi_i32(s, t, TCGV_LOW(s, arg1), arg2 - 32); tcg_gen_addi_i32(s, t, t, 32); tcg_gen_clz_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, arg1), t); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); tcg_temp_free_i32(s, t); } else { TCGv_i64 t0 = tcg_const_i64(s, arg2); tcg_gen_clz_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_ctz_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_ctz_i64) { tcg_gen_op3_i64(s, INDEX_op_ctz_i64, ret, arg1, arg2); } else if (TCG_TARGET_HAS_ctpop_i64 || TCG_TARGET_HAS_clz_i64) { TCGv_i64 z, t = tcg_temp_new_i64(s); if (TCG_TARGET_HAS_ctpop_i64) { tcg_gen_subi_i64(s, t, arg1, 1); tcg_gen_andc_i64(s, t, t, arg1); tcg_gen_ctpop_i64(s, t, t); } else { /* Since all non-x86 hosts have clz(0) == 64, don't fight it. */ tcg_gen_neg_i64(s, t, arg1); tcg_gen_and_i64(s, t, t, arg1); tcg_gen_clzi_i64(s, t, t, 64); tcg_gen_xori_i64(s, t, t, 63); } z = tcg_constant_i64(s, 0); tcg_gen_movcond_i64(s, TCG_COND_EQ, ret, arg1, z, arg2, t); tcg_temp_free_i64(s, t); } else { gen_helper_ctz_i64(s, ret, arg1, arg2); } } void tcg_gen_ctzi_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, uint64_t arg2) { if (TCG_TARGET_REG_BITS == 32 && TCG_TARGET_HAS_ctz_i32 && arg2 <= 0xffffffffu) { TCGv_i32 t32 = tcg_temp_new_i32(s); tcg_gen_ctzi_i32(s, t32, TCGV_HIGH(s, arg1), arg2 - 32); tcg_gen_addi_i32(s, t32, t32, 32); tcg_gen_ctz_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), t32); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); tcg_temp_free_i32(s, t32); } else if (!TCG_TARGET_HAS_ctz_i64 && TCG_TARGET_HAS_ctpop_i64 && arg2 == 64) { /* This equivalence has the advantage of not requiring a fixup. */ TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_subi_i64(s, t, arg1, 1); tcg_gen_andc_i64(s, t, t, arg1); tcg_gen_ctpop_i64(s, ret, t); tcg_temp_free_i64(s, t); } else { TCGv_i64 t0 = tcg_const_i64(s, arg2); tcg_gen_ctz_i64(s, ret, arg1, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_clrsb_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg) { if (TCG_TARGET_HAS_clz_i64 || TCG_TARGET_HAS_clz_i32) { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_sari_i64(s, t, arg, 63); tcg_gen_xor_i64(s, t, t, arg); tcg_gen_clzi_i64(s, t, t, 64); tcg_gen_subi_i64(s, ret, t, 1); tcg_temp_free_i64(s, t); } else { gen_helper_clrsb_i64(s, ret, arg); } } void tcg_gen_ctpop_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1) { if (TCG_TARGET_HAS_ctpop_i64) { tcg_gen_op2_i64(s, INDEX_op_ctpop_i64, ret, arg1); } else if (TCG_TARGET_REG_BITS == 32 && TCG_TARGET_HAS_ctpop_i32) { tcg_gen_ctpop_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1)); tcg_gen_ctpop_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1)); tcg_gen_add_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, ret), TCGV_HIGH(s, ret)); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else { gen_helper_ctpop_i64(s, ret, arg1); } } void tcg_gen_rotl_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_rot_i64) { tcg_gen_op3_i64(s, INDEX_op_rotl_i64, ret, arg1, arg2); } else { TCGv_i64 t0, t1; t0 = tcg_temp_new_i64(s); t1 = tcg_temp_new_i64(s); tcg_gen_shl_i64(s, t0, arg1, arg2); tcg_gen_subfi_i64(s, t1, 64, arg2); tcg_gen_shr_i64(s, t1, arg1, t1); tcg_gen_or_i64(s, ret, t0, t1); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_rotli_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 64); /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else if (TCG_TARGET_HAS_rot_i64) { tcg_gen_rotl_i64(s, ret, arg1, tcg_constant_i64(s, arg2)); } else { TCGv_i64 t0, t1; t0 = tcg_temp_new_i64(s); t1 = tcg_temp_new_i64(s); tcg_gen_shli_i64(s, t0, arg1, arg2); tcg_gen_shri_i64(s, t1, arg1, 64 - arg2); tcg_gen_or_i64(s, ret, t0, t1); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_rotr_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_rot_i64) { tcg_gen_op3_i64(s, INDEX_op_rotr_i64, ret, arg1, arg2); } else { TCGv_i64 t0, t1; t0 = tcg_temp_new_i64(s); t1 = tcg_temp_new_i64(s); tcg_gen_shr_i64(s, t0, arg1, arg2); tcg_gen_subfi_i64(s, t1, 64, arg2); tcg_gen_shl_i64(s, t1, arg1, t1); tcg_gen_or_i64(s, ret, t0, t1); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_rotri_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, int64_t arg2) { tcg_debug_assert(arg2 >= 0 && arg2 < 64); /* some cases can be optimized here */ if (arg2 == 0) { tcg_gen_mov_i64(s, ret, arg1); } else { tcg_gen_rotli_i64(s, ret, arg1, 64 - arg2); } } void tcg_gen_deposit_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg1, TCGv_i64 arg2, unsigned int ofs, unsigned int len) { uint64_t mask; TCGv_i64 t1; tcg_debug_assert(ofs < 64); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 64); tcg_debug_assert(ofs + len <= 64); if (len == 64) { tcg_gen_mov_i64(s, ret, arg2); return; } if (TCG_TARGET_HAS_deposit_i64 && TCG_TARGET_deposit_i64_valid(ofs, len)) { tcg_gen_op5ii_i64(s, INDEX_op_deposit_i64, ret, arg1, arg2, ofs, len); return; } if (TCG_TARGET_REG_BITS == 32) { if (ofs >= 32) { tcg_gen_deposit_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1), TCGV_LOW(s, arg2), ofs - 32, len); tcg_gen_mov_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1)); return; } if (ofs + len <= 32) { tcg_gen_deposit_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg1), TCGV_LOW(s, arg2), ofs, len); tcg_gen_mov_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg1)); return; } } t1 = tcg_temp_new_i64(s); if (TCG_TARGET_HAS_extract2_i64) { if (ofs + len == 64) { tcg_gen_shli_i64(s, t1, arg1, len); tcg_gen_extract2_i64(s, ret, t1, arg2, len); goto done; } if (ofs == 0) { tcg_gen_extract2_i64(s, ret, arg1, arg2, len); tcg_gen_rotli_i64(s, ret, ret, len); goto done; } } mask = (1ull << len) - 1; if (ofs + len < 64) { tcg_gen_andi_i64(s, t1, arg2, mask); tcg_gen_shli_i64(s, t1, t1, ofs); } else { tcg_gen_shli_i64(s, t1, arg2, ofs); } tcg_gen_andi_i64(s, ret, arg1, ~(mask << ofs)); tcg_gen_or_i64(s, ret, ret, t1); done: tcg_temp_free_i64(s, t1); } void tcg_gen_deposit_z_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 64); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 64); tcg_debug_assert(ofs + len <= 64); if (ofs + len == 64) { tcg_gen_shli_i64(s, ret, arg, ofs); } else if (ofs == 0) { tcg_gen_andi_i64(s, ret, arg, (1ull << len) - 1); } else if (TCG_TARGET_HAS_deposit_i64 && TCG_TARGET_deposit_i64_valid(ofs, len)) { TCGv_i64 zero = tcg_constant_i64(s, 0); tcg_gen_op5ii_i64(s, INDEX_op_deposit_i64, ret, zero, arg, ofs, len); } else { if (TCG_TARGET_REG_BITS == 32) { if (ofs >= 32) { tcg_gen_deposit_z_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, arg), ofs - 32, len); tcg_gen_movi_i32(s, TCGV_LOW(s, ret), 0); return; } if (ofs + len <= 32) { tcg_gen_deposit_z_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg), ofs, len); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); return; } } /* To help two-operand hosts we prefer to zero-extend first, which allows ARG to stay live. */ switch (len) { case 32: if (TCG_TARGET_HAS_ext32u_i64) { tcg_gen_ext32u_i64(s, ret, arg); tcg_gen_shli_i64(s, ret, ret, ofs); return; } break; case 16: if (TCG_TARGET_HAS_ext16u_i64) { tcg_gen_ext16u_i64(s, ret, arg); tcg_gen_shli_i64(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i64) { tcg_gen_ext8u_i64(s, ret, arg); tcg_gen_shli_i64(s, ret, ret, ofs); return; } break; } /* Otherwise prefer zero-extension over AND for code size. */ switch (ofs + len) { case 32: if (TCG_TARGET_HAS_ext32u_i64) { tcg_gen_shli_i64(s, ret, arg, ofs); tcg_gen_ext32u_i64(s, ret, ret); return; } break; case 16: if (TCG_TARGET_HAS_ext16u_i64) { tcg_gen_shli_i64(s, ret, arg, ofs); tcg_gen_ext16u_i64(s, ret, ret); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i64) { tcg_gen_shli_i64(s, ret, arg, ofs); tcg_gen_ext8u_i64(s, ret, ret); return; } break; } tcg_gen_andi_i64(s, ret, arg, (1ull << len) - 1); tcg_gen_shli_i64(s, ret, ret, ofs); } } void tcg_gen_extract_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 64); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 64); tcg_debug_assert(ofs + len <= 64); /* Canonicalize certain special cases, even if extract is supported. */ if (ofs + len == 64) { tcg_gen_shri_i64(s, ret, arg, 64 - len); return; } if (ofs == 0) { tcg_gen_andi_i64(s, ret, arg, (1ull << len) - 1); return; } if (TCG_TARGET_REG_BITS == 32) { /* Look for a 32-bit extract within one of the two words. */ if (ofs >= 32) { tcg_gen_extract_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, arg), ofs - 32, len); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); return; } if (ofs + len <= 32) { tcg_gen_extract_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg), ofs, len); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); return; } /* The field is split across two words. One double-word shift is better than two double-word shifts. */ goto do_shift_and; } if (TCG_TARGET_HAS_extract_i64 && TCG_TARGET_extract_i64_valid(ofs, len)) { tcg_gen_op4ii_i64(s, INDEX_op_extract_i64, ret, arg, ofs, len); return; } /* Assume that zero-extension, if available, is cheaper than a shift. */ switch (ofs + len) { case 32: if (TCG_TARGET_HAS_ext32u_i64) { tcg_gen_ext32u_i64(s, ret, arg); tcg_gen_shri_i64(s, ret, ret, ofs); return; } break; case 16: if (TCG_TARGET_HAS_ext16u_i64) { tcg_gen_ext16u_i64(s, ret, arg); tcg_gen_shri_i64(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8u_i64) { tcg_gen_ext8u_i64(s, ret, arg); tcg_gen_shri_i64(s, ret, ret, ofs); return; } break; } /* ??? Ideally we'd know what values are available for immediate AND. Assume that 8 bits are available, plus the special cases of 16 and 32, so that we get ext8u, ext16u, and ext32u. */ switch (len) { case 1: case 2: case 3: case 4: case 5: case 6: case 7: case 8: case 16: case 32: do_shift_and: tcg_gen_shri_i64(s, ret, arg, ofs); tcg_gen_andi_i64(s, ret, ret, (1ull << len) - 1); break; default: tcg_gen_shli_i64(s, ret, arg, 64 - len - ofs); tcg_gen_shri_i64(s, ret, ret, 64 - len); break; } } void tcg_gen_sextract_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 arg, unsigned int ofs, unsigned int len) { tcg_debug_assert(ofs < 64); tcg_debug_assert(len > 0); tcg_debug_assert(len <= 64); tcg_debug_assert(ofs + len <= 64); /* Canonicalize certain special cases, even if sextract is supported. */ if (ofs + len == 64) { tcg_gen_sari_i64(s, ret, arg, 64 - len); return; } if (ofs == 0) { switch (len) { case 32: tcg_gen_ext32s_i64(s, ret, arg); return; case 16: tcg_gen_ext16s_i64(s, ret, arg); return; case 8: tcg_gen_ext8s_i64(s, ret, arg); return; } } if (TCG_TARGET_REG_BITS == 32) { /* Look for a 32-bit extract within one of the two words. */ if (ofs >= 32) { tcg_gen_sextract_i32(s, TCGV_LOW(s, ret), TCGV_HIGH(s, arg), ofs - 32, len); } else if (ofs + len <= 32) { tcg_gen_sextract_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg), ofs, len); } else if (ofs == 0) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, arg)); tcg_gen_sextract_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, arg), 0, len - 32); return; } else if (len > 32) { TCGv_i32 t = tcg_temp_new_i32(s); /* Extract the bits for the high word normally. */ tcg_gen_sextract_i32(s, t, TCGV_HIGH(s, arg), ofs + 32, len - 32); /* Shift the field down for the low part. */ tcg_gen_shri_i64(s, ret, arg, ofs); /* Overwrite the shift into the high part. */ tcg_gen_mov_i32(s, TCGV_HIGH(s, ret), t); tcg_temp_free_i32(s, t); return; } else { /* Shift the field down for the low part, such that the field sits at the MSB. */ tcg_gen_shri_i64(s, ret, arg, ofs + len - 32); /* Shift the field down from the MSB, sign extending. */ tcg_gen_sari_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, ret), 32 - len); } /* Sign-extend the field from 32 bits. */ tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); return; } if (TCG_TARGET_HAS_sextract_i64 && TCG_TARGET_extract_i64_valid(ofs, len)) { tcg_gen_op4ii_i64(s, INDEX_op_sextract_i64, ret, arg, ofs, len); return; } /* Assume that sign-extension, if available, is cheaper than a shift. */ switch (ofs + len) { case 32: if (TCG_TARGET_HAS_ext32s_i64) { tcg_gen_ext32s_i64(s, ret, arg); tcg_gen_sari_i64(s, ret, ret, ofs); return; } break; case 16: if (TCG_TARGET_HAS_ext16s_i64) { tcg_gen_ext16s_i64(s, ret, arg); tcg_gen_sari_i64(s, ret, ret, ofs); return; } break; case 8: if (TCG_TARGET_HAS_ext8s_i64) { tcg_gen_ext8s_i64(s, ret, arg); tcg_gen_sari_i64(s, ret, ret, ofs); return; } break; } switch (len) { case 32: if (TCG_TARGET_HAS_ext32s_i64) { tcg_gen_shri_i64(s, ret, arg, ofs); tcg_gen_ext32s_i64(s, ret, ret); return; } break; case 16: if (TCG_TARGET_HAS_ext16s_i64) { tcg_gen_shri_i64(s, ret, arg, ofs); tcg_gen_ext16s_i64(s, ret, ret); return; } break; case 8: if (TCG_TARGET_HAS_ext8s_i64) { tcg_gen_shri_i64(s, ret, arg, ofs); tcg_gen_ext8s_i64(s, ret, ret); return; } break; } tcg_gen_shli_i64(s, ret, arg, 64 - len - ofs); tcg_gen_sari_i64(s, ret, ret, 64 - len); } /* * Extract 64 bits from a 128-bit input, ah:al, starting from ofs. * Unlike tcg_gen_extract_i64 above, len is fixed at 64. */ void tcg_gen_extract2_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 al, TCGv_i64 ah, unsigned int ofs) { tcg_debug_assert(ofs <= 64); if (ofs == 0) { tcg_gen_mov_i64(s, ret, al); } else if (ofs == 64) { tcg_gen_mov_i64(s, ret, ah); } else if (al == ah) { tcg_gen_rotri_i64(s, ret, al, ofs); } else if (TCG_TARGET_HAS_extract2_i64) { tcg_gen_op4i_i64(s, INDEX_op_extract2_i64, ret, al, ah, ofs); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_shri_i64(s, t0, al, ofs); tcg_gen_deposit_i64(s, ret, t0, ah, 64 - ofs, ofs); tcg_temp_free_i64(s, t0); } } void tcg_gen_movcond_i64(TCGContext *s, TCGCond cond, TCGv_i64 ret, TCGv_i64 c1, TCGv_i64 c2, TCGv_i64 v1, TCGv_i64 v2) { if (cond == TCG_COND_ALWAYS) { tcg_gen_mov_i64(s, ret, v1); } else if (cond == TCG_COND_NEVER) { tcg_gen_mov_i64(s, ret, v2); } else if (TCG_TARGET_REG_BITS == 32) { TCGv_i32 t0 = tcg_temp_new_i32(s); TCGv_i32 t1 = tcg_temp_new_i32(s); tcg_gen_op6i_i32(s, INDEX_op_setcond2_i32, t0, TCGV_LOW(s, c1), TCGV_HIGH(s, c1), TCGV_LOW(s, c2), TCGV_HIGH(s, c2), cond); if (TCG_TARGET_HAS_movcond_i32) { tcg_gen_movi_i32(s, t1, 0); tcg_gen_movcond_i32(s, TCG_COND_NE, TCGV_LOW(s, ret), t0, t1, TCGV_LOW(s, v1), TCGV_LOW(s, v2)); tcg_gen_movcond_i32(s, TCG_COND_NE, TCGV_HIGH(s, ret), t0, t1, TCGV_HIGH(s, v1), TCGV_HIGH(s, v2)); } else { tcg_gen_neg_i32(s, t0, t0); tcg_gen_and_i32(s, t1, TCGV_LOW(s, v1), t0); tcg_gen_andc_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, v2), t0); tcg_gen_or_i32(s, TCGV_LOW(s, ret), TCGV_LOW(s, ret), t1); tcg_gen_and_i32(s, t1, TCGV_HIGH(s, v1), t0); tcg_gen_andc_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, v2), t0); tcg_gen_or_i32(s, TCGV_HIGH(s, ret), TCGV_HIGH(s, ret), t1); } tcg_temp_free_i32(s, t0); tcg_temp_free_i32(s, t1); } else if (TCG_TARGET_HAS_movcond_i64) { tcg_gen_op6i_i64(s, INDEX_op_movcond_i64, ret, c1, c2, v1, v2, cond); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_setcond_i64(s, cond, t0, c1, c2); tcg_gen_neg_i64(s, t0, t0); tcg_gen_and_i64(s, t1, v1, t0); tcg_gen_andc_i64(s, ret, v2, t0); tcg_gen_or_i64(s, ret, ret, t1); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_add2_i64(TCGContext *s, TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 al, TCGv_i64 ah, TCGv_i64 bl, TCGv_i64 bh) { if (TCG_TARGET_HAS_add2_i64) { tcg_gen_op6_i64(s, INDEX_op_add2_i64, rl, rh, al, ah, bl, bh); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_add_i64(s, t0, al, bl); tcg_gen_setcond_i64(s, TCG_COND_LTU, t1, t0, al); tcg_gen_add_i64(s, rh, ah, bh); tcg_gen_add_i64(s, rh, rh, t1); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_sub2_i64(TCGContext *s, TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 al, TCGv_i64 ah, TCGv_i64 bl, TCGv_i64 bh) { if (TCG_TARGET_HAS_sub2_i64) { tcg_gen_op6_i64(s, INDEX_op_sub2_i64, rl, rh, al, ah, bl, bh); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); tcg_gen_sub_i64(s, t0, al, bl); tcg_gen_setcond_i64(s, TCG_COND_LTU, t1, al, bl); tcg_gen_sub_i64(s, rh, ah, bh); tcg_gen_sub_i64(s, rh, rh, t1); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); } } void tcg_gen_mulu2_i64(TCGContext *s, TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_mulu2_i64) { tcg_gen_op4_i64(s, INDEX_op_mulu2_i64, rl, rh, arg1, arg2); } else if (TCG_TARGET_HAS_muluh_i64) { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_op3_i64(s, INDEX_op_mul_i64, t, arg1, arg2); tcg_gen_op3_i64(s, INDEX_op_muluh_i64, rh, arg1, arg2); tcg_gen_mov_i64(s, rl, t); tcg_temp_free_i64(s, t); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_mul_i64(s, t0, arg1, arg2); gen_helper_muluh_i64(s, rh, arg1, arg2); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_muls2_i64(TCGContext *s, TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2) { if (TCG_TARGET_HAS_muls2_i64) { tcg_gen_op4_i64(s, INDEX_op_muls2_i64, rl, rh, arg1, arg2); } else if (TCG_TARGET_HAS_mulsh_i64) { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_op3_i64(s, INDEX_op_mul_i64, t, arg1, arg2); tcg_gen_op3_i64(s, INDEX_op_mulsh_i64, rh, arg1, arg2); tcg_gen_mov_i64(s, rl, t); tcg_temp_free_i64(s, t); } else if (TCG_TARGET_HAS_mulu2_i64 || TCG_TARGET_HAS_muluh_i64) { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); TCGv_i64 t3 = tcg_temp_new_i64(s); tcg_gen_mulu2_i64(s, t0, t1, arg1, arg2); /* Adjust for negative inputs. */ tcg_gen_sari_i64(s, t2, arg1, 63); tcg_gen_sari_i64(s, t3, arg2, 63); tcg_gen_and_i64(s, t2, t2, arg2); tcg_gen_and_i64(s, t3, t3, arg1); tcg_gen_sub_i64(s, rh, t1, t2); tcg_gen_sub_i64(s, rh, rh, t3); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); tcg_temp_free_i64(s, t3); } else { TCGv_i64 t0 = tcg_temp_new_i64(s); tcg_gen_mul_i64(s, t0, arg1, arg2); gen_helper_mulsh_i64(s, rh, arg1, arg2); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); } } void tcg_gen_mulsu2_i64(TCGContext *s, TCGv_i64 rl, TCGv_i64 rh, TCGv_i64 arg1, TCGv_i64 arg2) { TCGv_i64 t0 = tcg_temp_new_i64(s); TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); tcg_gen_mulu2_i64(s, t0, t1, arg1, arg2); /* Adjust for negative input for the signed arg1. */ tcg_gen_sari_i64(s, t2, arg1, 63); tcg_gen_and_i64(s, t2, t2, arg2); tcg_gen_sub_i64(s, rh, t1, t2); tcg_gen_mov_i64(s, rl, t0); tcg_temp_free_i64(s, t0); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); } void tcg_gen_smin_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 a, TCGv_i64 b) { tcg_gen_movcond_i64(s, TCG_COND_LT, ret, a, b, a, b); } void tcg_gen_umin_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 a, TCGv_i64 b) { tcg_gen_movcond_i64(s, TCG_COND_LTU, ret, a, b, a, b); } void tcg_gen_smax_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 a, TCGv_i64 b) { tcg_gen_movcond_i64(s, TCG_COND_LT, ret, a, b, b, a); } void tcg_gen_umax_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 a, TCGv_i64 b) { tcg_gen_movcond_i64(s, TCG_COND_LTU, ret, a, b, b, a); } void tcg_gen_abs_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 a) { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_sari_i64(s, t, a, 63); tcg_gen_xor_i64(s, ret, a, t); tcg_gen_sub_i64(s, ret, ret, t); tcg_temp_free_i64(s, t); } /* Size changing operations. */ void tcg_gen_extrl_i64_i32(TCGContext *s, TCGv_i32 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, ret, TCGV_LOW(s, arg)); } else if (TCG_TARGET_HAS_extrl_i64_i32) { tcg_gen_op2(s, INDEX_op_extrl_i64_i32, tcgv_i32_arg(s, ret), tcgv_i64_arg(s, arg)); } else { tcg_gen_mov_i32(s, ret, (TCGv_i32)arg); } } void tcg_gen_extrh_i64_i32(TCGContext *s, TCGv_i32 ret, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, ret, TCGV_HIGH(s, arg)); } else if (TCG_TARGET_HAS_extrh_i64_i32) { tcg_gen_op2(s, INDEX_op_extrh_i64_i32, tcgv_i32_arg(s, ret), tcgv_i64_arg(s, arg)); } else { TCGv_i64 t = tcg_temp_new_i64(s); tcg_gen_shri_i64(s, t, arg, 32); tcg_gen_mov_i32(s, ret, (TCGv_i32)t); tcg_temp_free_i64(s, t); } } void tcg_gen_extu_i32_i64(TCGContext *s, TCGv_i64 ret, TCGv_i32 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), arg); tcg_gen_movi_i32(s, TCGV_HIGH(s, ret), 0); } else { tcg_gen_op2(s, INDEX_op_extu_i32_i64, tcgv_i64_arg(s, ret), tcgv_i32_arg(s, arg)); } } void tcg_gen_ext_i32_i64(TCGContext *s, TCGv_i64 ret, TCGv_i32 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, TCGV_LOW(s, ret), arg); tcg_gen_sari_i32(s, TCGV_HIGH(s, ret), TCGV_LOW(s, ret), 31); } else { tcg_gen_op2(s, INDEX_op_ext_i32_i64, tcgv_i64_arg(s, ret), tcgv_i32_arg(s, arg)); } } void tcg_gen_concat_i32_i64(TCGContext *s, TCGv_i64 dest, TCGv_i32 low, TCGv_i32 high) { TCGv_i64 tmp; if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, TCGV_LOW(s, dest), low); tcg_gen_mov_i32(s, TCGV_HIGH(s, dest), high); return; } tmp = tcg_temp_new_i64(s); /* These extensions are only needed for type correctness. We may be able to do better given target specific information. */ tcg_gen_extu_i32_i64(s, tmp, high); tcg_gen_extu_i32_i64(s, dest, low); /* If deposit is available, use it. Otherwise use the extra knowledge that we have of the zero-extensions above. */ if (TCG_TARGET_HAS_deposit_i64 && TCG_TARGET_deposit_i64_valid(32, 32)) { tcg_gen_deposit_i64(s, dest, dest, tmp, 32, 32); } else { tcg_gen_shli_i64(s, tmp, tmp, 32); tcg_gen_or_i64(s, dest, dest, tmp); } tcg_temp_free_i64(s, tmp); } void tcg_gen_extr_i64_i32(TCGContext *s, TCGv_i32 lo, TCGv_i32 hi, TCGv_i64 arg) { if (TCG_TARGET_REG_BITS == 32) { tcg_gen_mov_i32(s, lo, TCGV_LOW(s, arg)); tcg_gen_mov_i32(s, hi, TCGV_HIGH(s, arg)); } else { tcg_gen_extrl_i64_i32(s, lo, arg); tcg_gen_extrh_i64_i32(s, hi, arg); } } void tcg_gen_extr32_i64(TCGContext *s, TCGv_i64 lo, TCGv_i64 hi, TCGv_i64 arg) { tcg_gen_ext32u_i64(s, lo, arg); tcg_gen_shri_i64(s, hi, arg, 32); } /* QEMU specific operations. */ void tcg_gen_exit_tb(TCGContext *s, TranslationBlock *tb, unsigned idx) { uintptr_t val = (uintptr_t)tb + idx; if (tb == NULL) { tcg_debug_assert(idx == 0); } else if (idx <= TB_EXIT_IDXMAX) { #ifdef CONFIG_DEBUG_TCG /* This is an exit following a goto_tb. Verify that we have seen this numbered exit before, via tcg_gen_goto_tb. */ tcg_debug_assert(tcg_ctx->goto_tb_issue_mask & (1 << idx)); #endif /* When not chaining, exit without indicating a link. */ if (qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { val = 0; } } else { /* This is an exit via the exitreq label. */ tcg_debug_assert(idx == TB_EXIT_REQUESTED); } tcg_gen_op1i(s, INDEX_op_exit_tb, val); } void tcg_gen_goto_tb(TCGContext *s, unsigned idx) { /* We only support two chained exits. */ tcg_debug_assert(idx <= TB_EXIT_IDXMAX); #ifdef CONFIG_DEBUG_TCG /* Verify that we havn't seen this numbered exit before. */ tcg_debug_assert((s->goto_tb_issue_mask & (1 << idx)) == 0); s->goto_tb_issue_mask |= 1 << idx; #endif /* When not chaining, we simply fall through to the "fallback" exit. */ if (!qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { tcg_gen_op1i(s, INDEX_op_goto_tb, idx); } } void tcg_gen_lookup_and_goto_ptr(TCGContext *s) { if (TCG_TARGET_HAS_goto_ptr && !qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) { TCGv_ptr ptr = tcg_temp_new_ptr(s); gen_helper_lookup_tb_ptr(s, ptr, s->tcg_env); tcg_gen_op1i(s, INDEX_op_goto_ptr, tcgv_ptr_arg(s, ptr)); tcg_temp_free_ptr(s, ptr); } else { tcg_gen_exit_tb(s, NULL, 0); } } static inline MemOp tcg_canonicalize_memop(MemOp op, bool is64, bool st) { /* Trigger the asserts within as early as possible. */ (void)get_alignment_bits(op); switch (op & MO_SIZE) { case MO_8: op &= ~MO_BSWAP; break; case MO_16: break; case MO_32: if (!is64) { op &= ~MO_SIGN; } break; case MO_64: if (!is64) { tcg_abort(); } break; } if (st) { op &= ~MO_SIGN; } return op; } static void gen_ldst_i32(TCGContext *s, TCGOpcode opc, TCGv_i32 val, TCGv addr, MemOp memop, TCGArg idx) { TCGMemOpIdx oi = make_memop_idx(memop, idx); #if TARGET_LONG_BITS == 32 tcg_gen_op3i_i32(s, opc, val, addr, oi); #else if (TCG_TARGET_REG_BITS == 32) { tcg_gen_op4i_i32(s, opc, val, TCGV_LOW(s, addr), TCGV_HIGH(s, addr), oi); } else { tcg_gen_op3(s, opc, tcgv_i32_arg(s, val), tcgv_i64_arg(s, addr), oi); } #endif } static void gen_ldst_i64(TCGContext *s, TCGOpcode opc, TCGv_i64 val, TCGv addr, MemOp memop, TCGArg idx) { TCGMemOpIdx oi = make_memop_idx(memop, idx); #if TARGET_LONG_BITS == 32 if (TCG_TARGET_REG_BITS == 32) { tcg_gen_op4i_i32(s, opc, TCGV_LOW(s, val), TCGV_HIGH(s, val), addr, oi); } else { tcg_gen_op3(s, opc, tcgv_i64_arg(s, val), tcgv_i32_arg(s, addr), oi); } #else if (TCG_TARGET_REG_BITS == 32) { tcg_gen_op5i_i32(s, opc, TCGV_LOW(s, val), TCGV_HIGH(s, val), TCGV_LOW(s, addr), TCGV_HIGH(s, addr), oi); } else { tcg_gen_op3i_i64(s, opc, val, addr, oi); } #endif } // Unicorn engine // check if the last memory access was invalid // if so, we jump to the block epilogue to quit immediately. void check_exit_request(TCGContext *tcg_ctx) { TCGv_i32 flag; flag = tcg_temp_new_i32(tcg_ctx); tcg_gen_ld_i32(tcg_ctx, flag, tcg_ctx->cpu_env, offsetof(CPUState, tcg_exit_req) - offsetof(ArchCPU, env)); tcg_gen_brcondi_i32(tcg_ctx, TCG_COND_NE, flag, 0, tcg_ctx->exitreq_label); tcg_temp_free_i32(tcg_ctx, flag); } static void tcg_gen_req_mo(TCGContext *s, TCGBar type) { #ifdef TCG_GUEST_DEFAULT_MO type &= TCG_GUEST_DEFAULT_MO; #endif type &= ~TCG_TARGET_DEFAULT_MO; if (type) { tcg_gen_mb(s, type | TCG_BAR_SC); } } void tcg_gen_qemu_ld_i32(struct uc_struct *uc, TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) { MemOp orig_memop; TCGContext *tcg_ctx = uc->tcg_ctx; tcg_gen_req_mo(tcg_ctx, TCG_MO_LD_LD | TCG_MO_ST_LD); memop = tcg_canonicalize_memop(memop, 0, 0); orig_memop = memop; if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { memop &= ~MO_BSWAP; /* The bswap primitive requires zero-extended input. */ if ((memop & MO_SSIZE) == MO_SW) { memop &= ~MO_SIGN; } } gen_ldst_i32(tcg_ctx, INDEX_op_qemu_ld_i32, val, addr, memop, idx); if ((orig_memop ^ memop) & MO_BSWAP) { switch (orig_memop & MO_SIZE) { case MO_16: tcg_gen_bswap16_i32(tcg_ctx, val, val); if (orig_memop & MO_SIGN) { tcg_gen_ext16s_i32(tcg_ctx, val, val); } break; case MO_32: tcg_gen_bswap32_i32(tcg_ctx, val, val); break; default: g_assert_not_reached(); } } check_exit_request(tcg_ctx); } void tcg_gen_qemu_st_i32(struct uc_struct *uc, TCGv_i32 val, TCGv addr, TCGArg idx, MemOp memop) { TCGv_i32 swap = NULL; TCGContext *tcg_ctx = uc->tcg_ctx; tcg_gen_req_mo(tcg_ctx, TCG_MO_LD_ST | TCG_MO_ST_ST); memop = tcg_canonicalize_memop(memop, 0, 1); if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { swap = tcg_temp_new_i32(tcg_ctx); switch (memop & MO_SIZE) { case MO_16: tcg_gen_ext16u_i32(tcg_ctx, swap, val); tcg_gen_bswap16_i32(tcg_ctx, swap, swap); break; case MO_32: tcg_gen_bswap32_i32(tcg_ctx, swap, val); break; default: g_assert_not_reached(); } val = swap; memop &= ~MO_BSWAP; } if (TCG_TARGET_HAS_qemu_st8_i32 && (memop & MO_SIZE) == MO_8) { gen_ldst_i32(tcg_ctx, INDEX_op_qemu_st8_i32, val, addr, memop, idx); } else { gen_ldst_i32(tcg_ctx, INDEX_op_qemu_st_i32, val, addr, memop, idx); } if (swap) { tcg_temp_free_i32(tcg_ctx, swap); } check_exit_request(tcg_ctx); } void tcg_gen_qemu_ld_i64(struct uc_struct *uc, TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) { TCGContext *tcg_ctx = uc->tcg_ctx; MemOp orig_memop; if (TCG_TARGET_REG_BITS == 32 && (memop & MO_SIZE) < MO_64) { tcg_gen_qemu_ld_i32(uc, TCGV_LOW(tcg_ctx, val), addr, idx, memop); if (memop & MO_SIGN) { tcg_gen_sari_i32(tcg_ctx, TCGV_HIGH(tcg_ctx, val), TCGV_LOW(tcg_ctx, val), 31); } else { tcg_gen_movi_i32(tcg_ctx, TCGV_HIGH(tcg_ctx, val), 0); } check_exit_request(tcg_ctx); return; } tcg_gen_req_mo(tcg_ctx, TCG_MO_LD_LD | TCG_MO_ST_LD); memop = tcg_canonicalize_memop(memop, 1, 0); orig_memop = memop; if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { memop &= ~MO_BSWAP; /* The bswap primitive requires zero-extended input. */ if ((memop & MO_SIGN) && (memop & MO_SIZE) < MO_64) { memop &= ~MO_SIGN; } } gen_ldst_i64(tcg_ctx, INDEX_op_qemu_ld_i64, val, addr, memop, idx); if ((orig_memop ^ memop) & MO_BSWAP) { switch (orig_memop & MO_SIZE) { case MO_16: tcg_gen_bswap16_i64(tcg_ctx, val, val); if (orig_memop & MO_SIGN) { tcg_gen_ext16s_i64(tcg_ctx, val, val); } break; case MO_32: tcg_gen_bswap32_i64(tcg_ctx, val, val); if (orig_memop & MO_SIGN) { tcg_gen_ext32s_i64(tcg_ctx, val, val); } break; case MO_64: tcg_gen_bswap64_i64(tcg_ctx, val, val); break; default: g_assert_not_reached(); } } check_exit_request(tcg_ctx); } void tcg_gen_qemu_st_i64(struct uc_struct *uc, TCGv_i64 val, TCGv addr, TCGArg idx, MemOp memop) { TCGContext *tcg_ctx = uc->tcg_ctx; TCGv_i64 swap = NULL; if (TCG_TARGET_REG_BITS == 32 && (memop & MO_SIZE) < MO_64) { tcg_gen_qemu_st_i32(uc, TCGV_LOW(tcg_ctx, val), addr, idx, memop); check_exit_request(tcg_ctx); return; } tcg_gen_req_mo(tcg_ctx, TCG_MO_LD_ST | TCG_MO_ST_ST); memop = tcg_canonicalize_memop(memop, 1, 1); if (!TCG_TARGET_HAS_MEMORY_BSWAP && (memop & MO_BSWAP)) { swap = tcg_temp_new_i64(tcg_ctx); switch (memop & MO_SIZE) { case MO_16: tcg_gen_ext16u_i64(tcg_ctx, swap, val); tcg_gen_bswap16_i64(tcg_ctx, swap, swap); break; case MO_32: tcg_gen_ext32u_i64(tcg_ctx, swap, val); tcg_gen_bswap32_i64(tcg_ctx, swap, swap); break; case MO_64: tcg_gen_bswap64_i64(tcg_ctx, swap, val); break; default: g_assert_not_reached(); } val = swap; memop &= ~MO_BSWAP; } gen_ldst_i64(tcg_ctx, INDEX_op_qemu_st_i64, val, addr, memop, idx); if (swap) { tcg_temp_free_i64(tcg_ctx, swap); } check_exit_request(tcg_ctx); } static void tcg_gen_ext_i32(TCGContext *s, TCGv_i32 ret, TCGv_i32 val, MemOp opc) { switch (opc & MO_SSIZE) { case MO_SB: tcg_gen_ext8s_i32(s, ret, val); break; case MO_UB: tcg_gen_ext8u_i32(s, ret, val); break; case MO_SW: tcg_gen_ext16s_i32(s, ret, val); break; case MO_UW: tcg_gen_ext16u_i32(s, ret, val); break; default: tcg_gen_mov_i32(s, ret, val); break; } } static void tcg_gen_ext_i64(TCGContext *s, TCGv_i64 ret, TCGv_i64 val, MemOp opc) { switch (opc & MO_SSIZE) { case MO_SB: tcg_gen_ext8s_i64(s, ret, val); break; case MO_UB: tcg_gen_ext8u_i64(s, ret, val); break; case MO_SW: tcg_gen_ext16s_i64(s, ret, val); break; case MO_UW: tcg_gen_ext16u_i64(s, ret, val); break; case MO_SL: tcg_gen_ext32s_i64(s, ret, val); break; case MO_UL: tcg_gen_ext32u_i64(s, ret, val); break; default: tcg_gen_mov_i64(s, ret, val); break; } } #ifdef CONFIG_SOFTMMU typedef void (*gen_atomic_cx_i32)(TCGContext *, TCGv_i32, TCGv_env, TCGv, TCGv_i32, TCGv_i32, TCGv_i32); typedef void (*gen_atomic_cx_i64)(TCGContext *, TCGv_i64, TCGv_env, TCGv, TCGv_i64, TCGv_i64, TCGv_i32); typedef void (*gen_atomic_op_i32)(TCGContext *, TCGv_i32, TCGv_env, TCGv, TCGv_i32, TCGv_i32); typedef void (*gen_atomic_op_i64)(TCGContext *, TCGv_i64, TCGv_env, TCGv, TCGv_i64, TCGv_i32); #else typedef void (*gen_atomic_cx_i32)(TCGContext *, TCGv_i32, TCGv_env, TCGv, TCGv_i32, TCGv_i32); typedef void (*gen_atomic_cx_i64)(TCGContext *, TCGv_i64, TCGv_env, TCGv, TCGv_i64, TCGv_i64); typedef void (*gen_atomic_op_i32)(TCGContext *, TCGv_i32, TCGv_env, TCGv, TCGv_i32); typedef void (*gen_atomic_op_i64)(TCGContext *, TCGv_i64, TCGv_env, TCGv, TCGv_i64); #endif #ifdef CONFIG_ATOMIC64 # define WITH_ATOMIC64(X) X, #else # define WITH_ATOMIC64(X) #endif static void * const table_cmpxchg[16] = { [MO_8] = gen_helper_atomic_cmpxchgb, [MO_16 | MO_LE] = gen_helper_atomic_cmpxchgw_le, [MO_16 | MO_BE] = gen_helper_atomic_cmpxchgw_be, [MO_32 | MO_LE] = gen_helper_atomic_cmpxchgl_le, [MO_32 | MO_BE] = gen_helper_atomic_cmpxchgl_be, WITH_ATOMIC64([MO_64 | MO_LE] = gen_helper_atomic_cmpxchgq_le) WITH_ATOMIC64([MO_64 | MO_BE] = gen_helper_atomic_cmpxchgq_be) }; void tcg_gen_atomic_cmpxchg_i32(TCGContext *s, TCGv_i32 retv, TCGv addr, TCGv_i32 cmpv, TCGv_i32 newv, TCGArg idx, MemOp memop) { memop = tcg_canonicalize_memop(memop, 0, 0); if (!(s->tb_cflags & CF_PARALLEL)) { TCGv_i32 t1 = tcg_temp_new_i32(s); TCGv_i32 t2 = tcg_temp_new_i32(s); tcg_gen_ext_i32(s, t2, cmpv, memop & MO_SIZE); tcg_gen_qemu_ld_i32(s->uc, t1, addr, idx, memop & ~MO_SIGN); tcg_gen_movcond_i32(s, TCG_COND_EQ, t2, t1, t2, newv, t1); tcg_gen_qemu_st_i32(s->uc, t2, addr, idx, memop); tcg_temp_free_i32(s, t2); if (memop & MO_SIGN) { tcg_gen_ext_i32(s, retv, t1, memop); } else { tcg_gen_mov_i32(s, retv, t1); } tcg_temp_free_i32(s, t1); } else { gen_atomic_cx_i32 gen; gen = table_cmpxchg[memop & (MO_SIZE | MO_BSWAP)]; tcg_debug_assert(gen != NULL); #ifdef CONFIG_SOFTMMU { TCGMemOpIdx oi = make_memop_idx(memop & ~MO_SIGN, idx); gen(s, retv, s->cpu_env, addr, cmpv, newv, tcg_constant_i32(s, oi)); } #else gen(s, retv, s->tcg_env, addr, cmpv, newv); #endif if (memop & MO_SIGN) { tcg_gen_ext_i32(s, retv, retv, memop); } } } void tcg_gen_atomic_cmpxchg_i64(TCGContext *s, TCGv_i64 retv, TCGv addr, TCGv_i64 cmpv, TCGv_i64 newv, TCGArg idx, MemOp memop) { memop = tcg_canonicalize_memop(memop, 1, 0); if (!(s->tb_cflags & CF_PARALLEL)) { TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); tcg_gen_ext_i64(s, t2, cmpv, memop & MO_SIZE); tcg_gen_qemu_ld_i64(s->uc, t1, addr, idx, memop & ~MO_SIGN); tcg_gen_movcond_i64(s, TCG_COND_EQ, t2, t1, t2, newv, t1); tcg_gen_qemu_st_i64(s->uc, t2, addr, idx, memop); tcg_temp_free_i64(s, t2); if (memop & MO_SIGN) { tcg_gen_ext_i64(s, retv, t1, memop); } else { tcg_gen_mov_i64(s, retv, t1); } tcg_temp_free_i64(s, t1); } else if ((memop & MO_SIZE) == MO_64) { #ifdef CONFIG_ATOMIC64 gen_atomic_cx_i64 gen; gen = table_cmpxchg[memop & (MO_SIZE | MO_BSWAP)]; tcg_debug_assert(gen != NULL); #ifdef CONFIG_SOFTMMU { TCGMemOpIdx oi = make_memop_idx(memop, idx); gen(s, retv, s->cpu_env, addr, cmpv, newv, tcg_constant_i32(s, oi)); } #else gen(s, retv, s->tcg_env, addr, cmpv, newv); #endif #else gen_helper_exit_atomic(s, s->tcg_env); /* Produce a result, so that we have a well-formed opcode stream with respect to uses of the result in the (dead) code following. */ tcg_gen_movi_i64(s, retv, 0); #endif /* CONFIG_ATOMIC64 */ } else { TCGv_i32 c32 = tcg_temp_new_i32(s); TCGv_i32 n32 = tcg_temp_new_i32(s); TCGv_i32 r32 = tcg_temp_new_i32(s); tcg_gen_extrl_i64_i32(s, c32, cmpv); tcg_gen_extrl_i64_i32(s, n32, newv); tcg_gen_atomic_cmpxchg_i32(s, r32, addr, c32, n32, idx, memop & ~MO_SIGN); tcg_temp_free_i32(s, c32); tcg_temp_free_i32(s, n32); tcg_gen_extu_i32_i64(s, retv, r32); tcg_temp_free_i32(s, r32); if (memop & MO_SIGN) { tcg_gen_ext_i64(s, retv, retv, memop); } } } static void do_nonatomic_op_i32(TCGContext *s, TCGv_i32 ret, TCGv addr, TCGv_i32 val, TCGArg idx, MemOp memop, bool new_val, void (*gen)(TCGContext *, TCGv_i32, TCGv_i32, TCGv_i32)) { TCGv_i32 t1 = tcg_temp_new_i32(s); TCGv_i32 t2 = tcg_temp_new_i32(s); memop = tcg_canonicalize_memop(memop, 0, 0); tcg_gen_qemu_ld_i32(s->uc, t1, addr, idx, memop); tcg_gen_ext_i32(s, t2, val, memop); gen(s, t2, t1, t2); tcg_gen_qemu_st_i32(s->uc, t2, addr, idx, memop); tcg_gen_ext_i32(s, ret, (new_val ? t2 : t1), memop); tcg_temp_free_i32(s, t1); tcg_temp_free_i32(s, t2); } static void do_atomic_op_i32(TCGContext *s, TCGv_i32 ret, TCGv addr, TCGv_i32 val, TCGArg idx, MemOp memop, void * const table[]) { gen_atomic_op_i32 gen; memop = tcg_canonicalize_memop(memop, 0, 0); gen = table[memop & (MO_SIZE | MO_BSWAP)]; tcg_debug_assert(gen != NULL); #ifdef CONFIG_SOFTMMU { TCGMemOpIdx oi = make_memop_idx(memop & ~MO_SIGN, idx); gen(s, ret, s->cpu_env, addr, val, tcg_constant_i32(s, oi)); } #else gen(s, ret, s->tcg_env, addr, val); #endif if (memop & MO_SIGN) { tcg_gen_ext_i32(s, ret, ret, memop); } } static void do_nonatomic_op_i64(TCGContext *s, TCGv_i64 ret, TCGv addr, TCGv_i64 val, TCGArg idx, MemOp memop, bool new_val, void (*gen)(TCGContext *, TCGv_i64, TCGv_i64, TCGv_i64)) { TCGv_i64 t1 = tcg_temp_new_i64(s); TCGv_i64 t2 = tcg_temp_new_i64(s); memop = tcg_canonicalize_memop(memop, 1, 0); tcg_gen_qemu_ld_i64(s->uc, t1, addr, idx, memop); tcg_gen_ext_i64(s, t2, val, memop); gen(s, t2, t1, t2); tcg_gen_qemu_st_i64(s->uc, t2, addr, idx, memop); tcg_gen_ext_i64(s, ret, (new_val ? t2 : t1), memop); tcg_temp_free_i64(s, t1); tcg_temp_free_i64(s, t2); } static void do_atomic_op_i64(TCGContext *s, TCGv_i64 ret, TCGv addr, TCGv_i64 val, TCGArg idx, MemOp memop, void * const table[]) { memop = tcg_canonicalize_memop(memop, 1, 0); if ((memop & MO_SIZE) == MO_64) { #ifdef CONFIG_ATOMIC64 gen_atomic_op_i64 gen; gen = table[memop & (MO_SIZE | MO_BSWAP)]; tcg_debug_assert(gen != NULL); #ifdef CONFIG_SOFTMMU { TCGMemOpIdx oi = make_memop_idx(memop & ~MO_SIGN, idx); gen(s, ret, s->cpu_env, addr, val, tcg_constant_i32(s, oi)); } #else gen(s, ret, s->tcg_env, addr, val); #endif #else gen_helper_exit_atomic(s, s->tcg_env); /* Produce a result, so that we have a well-formed opcode stream with respect to uses of the result in the (dead) code following. */ tcg_gen_movi_i64(s, ret, 0); #endif /* CONFIG_ATOMIC64 */ } else { TCGv_i32 v32 = tcg_temp_new_i32(s); TCGv_i32 r32 = tcg_temp_new_i32(s); tcg_gen_extrl_i64_i32(s, v32, val); do_atomic_op_i32(s, r32, addr, v32, idx, memop & ~MO_SIGN, table); tcg_temp_free_i32(s, v32); tcg_gen_extu_i32_i64(s, ret, r32); tcg_temp_free_i32(s, r32); if (memop & MO_SIGN) { tcg_gen_ext_i64(s, ret, ret, memop); } } } #define GEN_ATOMIC_HELPER(NAME, OP, NEW) \ static void * const table_##NAME[16] = { \ [MO_8] = gen_helper_atomic_##NAME##b, \ [MO_16 | MO_LE] = gen_helper_atomic_##NAME##w_le, \ [MO_16 | MO_BE] = gen_helper_atomic_##NAME##w_be, \ [MO_32 | MO_LE] = gen_helper_atomic_##NAME##l_le, \ [MO_32 | MO_BE] = gen_helper_atomic_##NAME##l_be, \ WITH_ATOMIC64([MO_64 | MO_LE] = gen_helper_atomic_##NAME##q_le) \ WITH_ATOMIC64([MO_64 | MO_BE] = gen_helper_atomic_##NAME##q_be) \ }; \ void tcg_gen_atomic_##NAME##_i32 \ (TCGContext *s, TCGv_i32 ret, TCGv addr, TCGv_i32 val, TCGArg idx, MemOp memop) \ { \ if (s->tb_cflags & CF_PARALLEL) { \ do_atomic_op_i32(s, ret, addr, val, idx, memop, table_##NAME); \ } else { \ do_nonatomic_op_i32(s, ret, addr, val, idx, memop, NEW, \ tcg_gen_##OP##_i32); \ } \ } \ void tcg_gen_atomic_##NAME##_i64 \ (TCGContext *s, TCGv_i64 ret, TCGv addr, TCGv_i64 val, TCGArg idx, MemOp memop) \ { \ if (s->tb_cflags & CF_PARALLEL) { \ do_atomic_op_i64(s, ret, addr, val, idx, memop, table_##NAME); \ } else { \ do_nonatomic_op_i64(s, ret, addr, val, idx, memop, NEW, \ tcg_gen_##OP##_i64); \ } \ } GEN_ATOMIC_HELPER(fetch_add, add, 0) GEN_ATOMIC_HELPER(fetch_and, and, 0) GEN_ATOMIC_HELPER(fetch_or, or, 0) GEN_ATOMIC_HELPER(fetch_xor, xor, 0) GEN_ATOMIC_HELPER(fetch_smin, smin, 0) GEN_ATOMIC_HELPER(fetch_umin, umin, 0) GEN_ATOMIC_HELPER(fetch_smax, smax, 0) GEN_ATOMIC_HELPER(fetch_umax, umax, 0) GEN_ATOMIC_HELPER(add_fetch, add, 1) GEN_ATOMIC_HELPER(and_fetch, and, 1) GEN_ATOMIC_HELPER(or_fetch, or, 1) GEN_ATOMIC_HELPER(xor_fetch, xor, 1) GEN_ATOMIC_HELPER(smin_fetch, smin, 1) GEN_ATOMIC_HELPER(umin_fetch, umin, 1) GEN_ATOMIC_HELPER(smax_fetch, smax, 1) GEN_ATOMIC_HELPER(umax_fetch, umax, 1) static void tcg_gen_mov2_i32(TCGContext *s, TCGv_i32 r, TCGv_i32 a, TCGv_i32 b) { tcg_gen_mov_i32(s, r, b); } static void tcg_gen_mov2_i64(TCGContext *s, TCGv_i64 r, TCGv_i64 a, TCGv_i64 b) { tcg_gen_mov_i64(s, r, b); } GEN_ATOMIC_HELPER(xchg, mov2, 0) #undef GEN_ATOMIC_HELPER