qemu-e2k/target-lm32/translate.c

1230 lines
32 KiB
C
Raw Normal View History

/*
* LatticeMico32 main translation routines.
*
* Copyright (c) 2010 Michael Walle <michael@walle.cc>
*
* 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, see <http://www.gnu.org/licenses/>.
*/
#include "cpu.h"
#include "disas.h"
#include "helper.h"
#include "tcg-op.h"
#include "hw/lm32_pic.h"
#define GEN_HELPER 1
#include "helper.h"
#define DISAS_LM32 1
#if DISAS_LM32
# define LOG_DIS(...) qemu_log_mask(CPU_LOG_TB_IN_ASM, ## __VA_ARGS__)
#else
# define LOG_DIS(...) do { } while (0)
#endif
#define EXTRACT_FIELD(src, start, end) \
(((src) >> start) & ((1 << (end - start + 1)) - 1))
#define MEM_INDEX 0
static TCGv_ptr cpu_env;
static TCGv cpu_R[32];
static TCGv cpu_pc;
static TCGv cpu_ie;
static TCGv cpu_icc;
static TCGv cpu_dcc;
static TCGv cpu_cc;
static TCGv cpu_cfg;
static TCGv cpu_eba;
static TCGv cpu_dc;
static TCGv cpu_deba;
static TCGv cpu_bp[4];
static TCGv cpu_wp[4];
#include "gen-icount.h"
enum {
OP_FMT_RI,
OP_FMT_RR,
OP_FMT_CR,
OP_FMT_I
};
/* This is the state at translation time. */
typedef struct DisasContext {
CPULM32State *env;
target_ulong pc;
/* Decoder. */
int format;
uint32_t ir;
uint8_t opcode;
uint8_t r0, r1, r2, csr;
uint16_t imm5;
uint16_t imm16;
uint32_t imm26;
unsigned int delayed_branch;
unsigned int tb_flags, synced_flags; /* tb dependent flags. */
int is_jmp;
int nr_nops;
struct TranslationBlock *tb;
int singlestep_enabled;
} DisasContext;
static const char *regnames[] = {
"r0", "r1", "r2", "r3", "r4", "r5", "r6", "r7",
"r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
"r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
"r24", "r25", "r26/gp", "r27/fp", "r28/sp", "r29/ra",
"r30/ea", "r31/ba", "bp0", "bp1", "bp2", "bp3", "wp0",
"wp1", "wp2", "wp3"
};
static inline int zero_extend(unsigned int val, int width)
{
return val & ((1 << width) - 1);
}
static inline int sign_extend(unsigned int val, int width)
{
int sval;
/* LSL. */
val <<= 32 - width;
sval = val;
/* ASR. */
sval >>= 32 - width;
return sval;
}
static inline void t_gen_raise_exception(DisasContext *dc, uint32_t index)
{
TCGv_i32 tmp = tcg_const_i32(index);
gen_helper_raise_exception(cpu_env, tmp);
tcg_temp_free_i32(tmp);
}
static void gen_goto_tb(DisasContext *dc, int n, target_ulong dest)
{
TranslationBlock *tb;
tb = dc->tb;
if ((tb->pc & TARGET_PAGE_MASK) == (dest & TARGET_PAGE_MASK) &&
likely(!dc->singlestep_enabled)) {
tcg_gen_goto_tb(n);
tcg_gen_movi_tl(cpu_pc, dest);
tcg_gen_exit_tb((tcg_target_long)tb + n);
} else {
tcg_gen_movi_tl(cpu_pc, dest);
if (dc->singlestep_enabled) {
t_gen_raise_exception(dc, EXCP_DEBUG);
}
tcg_gen_exit_tb(0);
}
}
static void dec_add(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
if (dc->r0 == R_R0) {
if (dc->r1 == R_R0 && dc->imm16 == 0) {
LOG_DIS("nop\n");
} else {
LOG_DIS("mvi r%d, %d\n", dc->r1, sign_extend(dc->imm16, 16));
}
} else {
LOG_DIS("addi r%d, r%d, %d\n", dc->r1, dc->r0,
sign_extend(dc->imm16, 16));
}
} else {
LOG_DIS("add r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (dc->format == OP_FMT_RI) {
tcg_gen_addi_tl(cpu_R[dc->r1], cpu_R[dc->r0],
sign_extend(dc->imm16, 16));
} else {
tcg_gen_add_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_and(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("andi r%d, r%d, %d\n", dc->r1, dc->r0,
zero_extend(dc->imm16, 16));
} else {
LOG_DIS("and r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (dc->format == OP_FMT_RI) {
tcg_gen_andi_tl(cpu_R[dc->r1], cpu_R[dc->r0],
zero_extend(dc->imm16, 16));
} else {
if (dc->r0 == 0 && dc->r1 == 0 && dc->r2 == 0) {
tcg_gen_movi_tl(cpu_pc, dc->pc + 4);
gen_helper_hlt(cpu_env);
} else {
tcg_gen_and_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
}
static void dec_andhi(DisasContext *dc)
{
LOG_DIS("andhi r%d, r%d, %d\n", dc->r2, dc->r0, dc->imm16);
tcg_gen_andi_tl(cpu_R[dc->r1], cpu_R[dc->r0], (dc->imm16 << 16));
}
static void dec_b(DisasContext *dc)
{
if (dc->r0 == R_RA) {
LOG_DIS("ret\n");
} else if (dc->r0 == R_EA) {
LOG_DIS("eret\n");
} else if (dc->r0 == R_BA) {
LOG_DIS("bret\n");
} else {
LOG_DIS("b r%d\n", dc->r0);
}
/* restore IE.IE in case of an eret */
if (dc->r0 == R_EA) {
TCGv t0 = tcg_temp_new();
int l1 = gen_new_label();
tcg_gen_andi_tl(t0, cpu_ie, IE_EIE);
tcg_gen_ori_tl(cpu_ie, cpu_ie, IE_IE);
tcg_gen_brcondi_tl(TCG_COND_EQ, t0, IE_EIE, l1);
tcg_gen_andi_tl(cpu_ie, cpu_ie, ~IE_IE);
gen_set_label(l1);
tcg_temp_free(t0);
} else if (dc->r0 == R_BA) {
TCGv t0 = tcg_temp_new();
int l1 = gen_new_label();
tcg_gen_andi_tl(t0, cpu_ie, IE_BIE);
tcg_gen_ori_tl(cpu_ie, cpu_ie, IE_IE);
tcg_gen_brcondi_tl(TCG_COND_EQ, t0, IE_BIE, l1);
tcg_gen_andi_tl(cpu_ie, cpu_ie, ~IE_IE);
gen_set_label(l1);
tcg_temp_free(t0);
}
tcg_gen_mov_tl(cpu_pc, cpu_R[dc->r0]);
dc->is_jmp = DISAS_JUMP;
}
static void dec_bi(DisasContext *dc)
{
LOG_DIS("bi %d\n", sign_extend(dc->imm26 << 2, 26));
gen_goto_tb(dc, 0, dc->pc + (sign_extend(dc->imm26 << 2, 26)));
dc->is_jmp = DISAS_TB_JUMP;
}
static inline void gen_cond_branch(DisasContext *dc, int cond)
{
int l1;
l1 = gen_new_label();
tcg_gen_brcond_tl(cond, cpu_R[dc->r0], cpu_R[dc->r1], l1);
gen_goto_tb(dc, 0, dc->pc + 4);
gen_set_label(l1);
gen_goto_tb(dc, 1, dc->pc + (sign_extend(dc->imm16 << 2, 16)));
dc->is_jmp = DISAS_TB_JUMP;
}
static void dec_be(DisasContext *dc)
{
LOG_DIS("be r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16) * 4);
gen_cond_branch(dc, TCG_COND_EQ);
}
static void dec_bg(DisasContext *dc)
{
LOG_DIS("bg r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16 * 4));
gen_cond_branch(dc, TCG_COND_GT);
}
static void dec_bge(DisasContext *dc)
{
LOG_DIS("bge r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16) * 4);
gen_cond_branch(dc, TCG_COND_GE);
}
static void dec_bgeu(DisasContext *dc)
{
LOG_DIS("bgeu r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16) * 4);
gen_cond_branch(dc, TCG_COND_GEU);
}
static void dec_bgu(DisasContext *dc)
{
LOG_DIS("bgu r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16) * 4);
gen_cond_branch(dc, TCG_COND_GTU);
}
static void dec_bne(DisasContext *dc)
{
LOG_DIS("bne r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16) * 4);
gen_cond_branch(dc, TCG_COND_NE);
}
static void dec_call(DisasContext *dc)
{
LOG_DIS("call r%d\n", dc->r0);
tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4);
tcg_gen_mov_tl(cpu_pc, cpu_R[dc->r0]);
dc->is_jmp = DISAS_JUMP;
}
static void dec_calli(DisasContext *dc)
{
LOG_DIS("calli %d\n", sign_extend(dc->imm26, 26) * 4);
tcg_gen_movi_tl(cpu_R[R_RA], dc->pc + 4);
gen_goto_tb(dc, 0, dc->pc + (sign_extend(dc->imm26 << 2, 26)));
dc->is_jmp = DISAS_TB_JUMP;
}
static inline void gen_compare(DisasContext *dc, int cond)
{
int rX = (dc->format == OP_FMT_RR) ? dc->r2 : dc->r1;
int rY = (dc->format == OP_FMT_RR) ? dc->r0 : dc->r0;
int rZ = (dc->format == OP_FMT_RR) ? dc->r1 : -1;
if (dc->format == OP_FMT_RI) {
tcg_gen_setcondi_tl(cond, cpu_R[rX], cpu_R[rY],
sign_extend(dc->imm16, 16));
} else {
tcg_gen_setcond_tl(cond, cpu_R[rX], cpu_R[rY], cpu_R[rZ]);
}
}
static void dec_cmpe(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpei r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpe r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_EQ);
}
static void dec_cmpg(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpgi r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpg r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_GT);
}
static void dec_cmpge(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpgei r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpge r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_GE);
}
static void dec_cmpgeu(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpgeui r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpgeu r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_GEU);
}
static void dec_cmpgu(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpgui r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpgu r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_GTU);
}
static void dec_cmpne(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("cmpnei r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("cmpne r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
gen_compare(dc, TCG_COND_NE);
}
static void dec_divu(DisasContext *dc)
{
int l1;
LOG_DIS("divu r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
if (!(dc->env->features & LM32_FEATURE_DIVIDE)) {
cpu_abort(dc->env, "hardware divider is not available\n");
}
l1 = gen_new_label();
tcg_gen_brcondi_tl(TCG_COND_NE, cpu_R[dc->r1], 0, l1);
tcg_gen_movi_tl(cpu_pc, dc->pc);
t_gen_raise_exception(dc, EXCP_DIVIDE_BY_ZERO);
gen_set_label(l1);
tcg_gen_divu_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
static void dec_lb(DisasContext *dc)
{
TCGv t0;
LOG_DIS("lb r%d, (r%d+%d)\n", dc->r1, dc->r0, dc->imm16);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_ld8s(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_lbu(DisasContext *dc)
{
TCGv t0;
LOG_DIS("lbu r%d, (r%d+%d)\n", dc->r1, dc->r0, dc->imm16);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_ld8u(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_lh(DisasContext *dc)
{
TCGv t0;
LOG_DIS("lh r%d, (r%d+%d)\n", dc->r1, dc->r0, dc->imm16);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_ld16s(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_lhu(DisasContext *dc)
{
TCGv t0;
LOG_DIS("lhu r%d, (r%d+%d)\n", dc->r1, dc->r0, dc->imm16);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_ld16u(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_lw(DisasContext *dc)
{
TCGv t0;
LOG_DIS("lw r%d, (r%d+%d)\n", dc->r1, dc->r0, sign_extend(dc->imm16, 16));
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_ld32s(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_modu(DisasContext *dc)
{
int l1;
LOG_DIS("modu r%d, r%d, %d\n", dc->r2, dc->r0, dc->r1);
if (!(dc->env->features & LM32_FEATURE_DIVIDE)) {
cpu_abort(dc->env, "hardware divider is not available\n");
}
l1 = gen_new_label();
tcg_gen_brcondi_tl(TCG_COND_NE, cpu_R[dc->r1], 0, l1);
tcg_gen_movi_tl(cpu_pc, dc->pc);
t_gen_raise_exception(dc, EXCP_DIVIDE_BY_ZERO);
gen_set_label(l1);
tcg_gen_remu_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
static void dec_mul(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("muli r%d, r%d, %d\n", dc->r0, dc->r1,
sign_extend(dc->imm16, 16));
} else {
LOG_DIS("mul r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (!(dc->env->features & LM32_FEATURE_MULTIPLY)) {
cpu_abort(dc->env, "hardware multiplier is not available\n");
}
if (dc->format == OP_FMT_RI) {
tcg_gen_muli_tl(cpu_R[dc->r1], cpu_R[dc->r0],
sign_extend(dc->imm16, 16));
} else {
tcg_gen_mul_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_nor(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("nori r%d, r%d, %d\n", dc->r0, dc->r1,
zero_extend(dc->imm16, 16));
} else {
LOG_DIS("nor r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (dc->format == OP_FMT_RI) {
TCGv t0 = tcg_temp_new();
tcg_gen_movi_tl(t0, zero_extend(dc->imm16, 16));
tcg_gen_nor_tl(cpu_R[dc->r1], cpu_R[dc->r0], t0);
tcg_temp_free(t0);
} else {
tcg_gen_nor_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_or(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("ori r%d, r%d, %d\n", dc->r1, dc->r0,
zero_extend(dc->imm16, 16));
} else {
if (dc->r1 == R_R0) {
LOG_DIS("mv r%d, r%d\n", dc->r2, dc->r0);
} else {
LOG_DIS("or r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
}
if (dc->format == OP_FMT_RI) {
tcg_gen_ori_tl(cpu_R[dc->r1], cpu_R[dc->r0],
zero_extend(dc->imm16, 16));
} else {
tcg_gen_or_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_orhi(DisasContext *dc)
{
if (dc->r0 == R_R0) {
LOG_DIS("mvhi r%d, %d\n", dc->r1, dc->imm16);
} else {
LOG_DIS("orhi r%d, r%d, %d\n", dc->r1, dc->r0, dc->imm16);
}
tcg_gen_ori_tl(cpu_R[dc->r1], cpu_R[dc->r0], (dc->imm16 << 16));
}
static void dec_scall(DisasContext *dc)
{
if (dc->imm5 == 7) {
LOG_DIS("scall\n");
} else if (dc->imm5 == 2) {
LOG_DIS("break\n");
} else {
cpu_abort(dc->env, "invalid opcode\n");
}
if (dc->imm5 == 7) {
tcg_gen_movi_tl(cpu_pc, dc->pc);
t_gen_raise_exception(dc, EXCP_SYSTEMCALL);
} else {
tcg_gen_movi_tl(cpu_pc, dc->pc);
t_gen_raise_exception(dc, EXCP_BREAKPOINT);
}
}
static void dec_rcsr(DisasContext *dc)
{
LOG_DIS("rcsr r%d, %d\n", dc->r2, dc->csr);
switch (dc->csr) {
case CSR_IE:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_ie);
break;
case CSR_IM:
gen_helper_rcsr_im(cpu_R[dc->r2], cpu_env);
break;
case CSR_IP:
gen_helper_rcsr_ip(cpu_R[dc->r2], cpu_env);
break;
case CSR_CC:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_cc);
break;
case CSR_CFG:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_cfg);
break;
case CSR_EBA:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_eba);
break;
case CSR_DC:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_dc);
break;
case CSR_DEBA:
tcg_gen_mov_tl(cpu_R[dc->r2], cpu_deba);
break;
case CSR_JTX:
gen_helper_rcsr_jtx(cpu_R[dc->r2], cpu_env);
break;
case CSR_JRX:
gen_helper_rcsr_jrx(cpu_R[dc->r2], cpu_env);
break;
case CSR_ICC:
case CSR_DCC:
case CSR_BP0:
case CSR_BP1:
case CSR_BP2:
case CSR_BP3:
case CSR_WP0:
case CSR_WP1:
case CSR_WP2:
case CSR_WP3:
cpu_abort(dc->env, "invalid read access csr=%x\n", dc->csr);
break;
default:
cpu_abort(dc->env, "read_csr: unknown csr=%x\n", dc->csr);
break;
}
}
static void dec_sb(DisasContext *dc)
{
TCGv t0;
LOG_DIS("sb (r%d+%d), r%d\n", dc->r0, dc->imm16, dc->r1);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_st8(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_sextb(DisasContext *dc)
{
LOG_DIS("sextb r%d, r%d\n", dc->r2, dc->r0);
if (!(dc->env->features & LM32_FEATURE_SIGN_EXTEND)) {
cpu_abort(dc->env, "hardware sign extender is not available\n");
}
tcg_gen_ext8s_tl(cpu_R[dc->r2], cpu_R[dc->r0]);
}
static void dec_sexth(DisasContext *dc)
{
LOG_DIS("sexth r%d, r%d\n", dc->r2, dc->r0);
if (!(dc->env->features & LM32_FEATURE_SIGN_EXTEND)) {
cpu_abort(dc->env, "hardware sign extender is not available\n");
}
tcg_gen_ext16s_tl(cpu_R[dc->r2], cpu_R[dc->r0]);
}
static void dec_sh(DisasContext *dc)
{
TCGv t0;
LOG_DIS("sh (r%d+%d), r%d\n", dc->r0, dc->imm16, dc->r1);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_st16(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_sl(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("sli r%d, r%d, %d\n", dc->r1, dc->r0, dc->imm5);
} else {
LOG_DIS("sl r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (!(dc->env->features & LM32_FEATURE_SHIFT)) {
cpu_abort(dc->env, "hardware shifter is not available\n");
}
if (dc->format == OP_FMT_RI) {
tcg_gen_shli_tl(cpu_R[dc->r1], cpu_R[dc->r0], dc->imm5);
} else {
TCGv t0 = tcg_temp_new();
tcg_gen_andi_tl(t0, cpu_R[dc->r1], 0x1f);
tcg_gen_shl_tl(cpu_R[dc->r2], cpu_R[dc->r0], t0);
tcg_temp_free(t0);
}
}
static void dec_sr(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("sri r%d, r%d, %d\n", dc->r1, dc->r0, dc->imm5);
} else {
LOG_DIS("sr r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (!(dc->env->features & LM32_FEATURE_SHIFT)) {
if (dc->format == OP_FMT_RI) {
/* TODO: check r1 == 1 during runtime */
} else {
if (dc->imm5 != 1) {
cpu_abort(dc->env, "hardware shifter is not available\n");
}
}
}
if (dc->format == OP_FMT_RI) {
tcg_gen_sari_tl(cpu_R[dc->r1], cpu_R[dc->r0], dc->imm5);
} else {
TCGv t0 = tcg_temp_new();
tcg_gen_andi_tl(t0, cpu_R[dc->r1], 0x1f);
tcg_gen_sar_tl(cpu_R[dc->r2], cpu_R[dc->r0], t0);
tcg_temp_free(t0);
}
}
static void dec_sru(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("srui r%d, r%d, %d\n", dc->r1, dc->r0, dc->imm5);
} else {
LOG_DIS("sru r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (!(dc->env->features & LM32_FEATURE_SHIFT)) {
if (dc->format == OP_FMT_RI) {
/* TODO: check r1 == 1 during runtime */
} else {
if (dc->imm5 != 1) {
cpu_abort(dc->env, "hardware shifter is not available\n");
}
}
}
if (dc->format == OP_FMT_RI) {
tcg_gen_shri_tl(cpu_R[dc->r1], cpu_R[dc->r0], dc->imm5);
} else {
TCGv t0 = tcg_temp_new();
tcg_gen_andi_tl(t0, cpu_R[dc->r1], 0x1f);
tcg_gen_shr_tl(cpu_R[dc->r2], cpu_R[dc->r0], t0);
tcg_temp_free(t0);
}
}
static void dec_sub(DisasContext *dc)
{
LOG_DIS("sub r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
tcg_gen_sub_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
static void dec_sw(DisasContext *dc)
{
TCGv t0;
LOG_DIS("sw (r%d+%d), r%d\n", dc->r0, sign_extend(dc->imm16, 16), dc->r1);
t0 = tcg_temp_new();
tcg_gen_addi_tl(t0, cpu_R[dc->r0], sign_extend(dc->imm16, 16));
tcg_gen_qemu_st32(cpu_R[dc->r1], t0, MEM_INDEX);
tcg_temp_free(t0);
}
static void dec_user(DisasContext *dc)
{
LOG_DIS("user");
cpu_abort(dc->env, "user insn undefined\n");
}
static void dec_wcsr(DisasContext *dc)
{
int no;
LOG_DIS("wcsr r%d, %d\n", dc->r1, dc->csr);
switch (dc->csr) {
case CSR_IE:
tcg_gen_mov_tl(cpu_ie, cpu_R[dc->r1]);
tcg_gen_movi_tl(cpu_pc, dc->pc + 4);
dc->is_jmp = DISAS_UPDATE;
break;
case CSR_IM:
/* mark as an io operation because it could cause an interrupt */
if (use_icount) {
gen_io_start();
}
gen_helper_wcsr_im(cpu_env, cpu_R[dc->r1]);
tcg_gen_movi_tl(cpu_pc, dc->pc + 4);
if (use_icount) {
gen_io_end();
}
dc->is_jmp = DISAS_UPDATE;
break;
case CSR_IP:
/* mark as an io operation because it could cause an interrupt */
if (use_icount) {
gen_io_start();
}
gen_helper_wcsr_ip(cpu_env, cpu_R[dc->r1]);
tcg_gen_movi_tl(cpu_pc, dc->pc + 4);
if (use_icount) {
gen_io_end();
}
dc->is_jmp = DISAS_UPDATE;
break;
case CSR_ICC:
/* TODO */
break;
case CSR_DCC:
/* TODO */
break;
case CSR_EBA:
tcg_gen_mov_tl(cpu_eba, cpu_R[dc->r1]);
break;
case CSR_DEBA:
tcg_gen_mov_tl(cpu_deba, cpu_R[dc->r1]);
break;
case CSR_JTX:
gen_helper_wcsr_jtx(cpu_env, cpu_R[dc->r1]);
break;
case CSR_JRX:
gen_helper_wcsr_jrx(cpu_env, cpu_R[dc->r1]);
break;
case CSR_DC:
tcg_gen_mov_tl(cpu_dc, cpu_R[dc->r1]);
break;
case CSR_BP0:
case CSR_BP1:
case CSR_BP2:
case CSR_BP3:
no = dc->csr - CSR_BP0;
if (dc->env->num_bps <= no) {
cpu_abort(dc->env, "breakpoint #%i is not available\n", no);
}
tcg_gen_mov_tl(cpu_bp[no], cpu_R[dc->r1]);
break;
case CSR_WP0:
case CSR_WP1:
case CSR_WP2:
case CSR_WP3:
no = dc->csr - CSR_WP0;
if (dc->env->num_wps <= no) {
cpu_abort(dc->env, "watchpoint #%i is not available\n", no);
}
tcg_gen_mov_tl(cpu_wp[no], cpu_R[dc->r1]);
break;
case CSR_CC:
case CSR_CFG:
cpu_abort(dc->env, "invalid write access csr=%x\n", dc->csr);
break;
default:
cpu_abort(dc->env, "write_csr unknown csr=%x\n", dc->csr);
break;
}
}
static void dec_xnor(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("xnori r%d, r%d, %d\n", dc->r0, dc->r1,
zero_extend(dc->imm16, 16));
} else {
if (dc->r1 == R_R0) {
LOG_DIS("not r%d, r%d\n", dc->r2, dc->r0);
} else {
LOG_DIS("xnor r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
}
if (dc->format == OP_FMT_RI) {
tcg_gen_xori_tl(cpu_R[dc->r1], cpu_R[dc->r0],
zero_extend(dc->imm16, 16));
tcg_gen_not_tl(cpu_R[dc->r1], cpu_R[dc->r1]);
} else {
tcg_gen_eqv_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_xor(DisasContext *dc)
{
if (dc->format == OP_FMT_RI) {
LOG_DIS("xori r%d, r%d, %d\n", dc->r0, dc->r1,
zero_extend(dc->imm16, 16));
} else {
LOG_DIS("xor r%d, r%d, r%d\n", dc->r2, dc->r0, dc->r1);
}
if (dc->format == OP_FMT_RI) {
tcg_gen_xori_tl(cpu_R[dc->r1], cpu_R[dc->r0],
zero_extend(dc->imm16, 16));
} else {
tcg_gen_xor_tl(cpu_R[dc->r2], cpu_R[dc->r0], cpu_R[dc->r1]);
}
}
static void dec_ill(DisasContext *dc)
{
cpu_abort(dc->env, "unknown opcode 0x%02x\n", dc->opcode);
}
typedef void (*DecoderInfo)(DisasContext *dc);
static const DecoderInfo decinfo[] = {
dec_sru, dec_nor, dec_mul, dec_sh, dec_lb, dec_sr, dec_xor, dec_lh,
dec_and, dec_xnor, dec_lw, dec_lhu, dec_sb, dec_add, dec_or, dec_sl,
dec_lbu, dec_be, dec_bg, dec_bge, dec_bgeu, dec_bgu, dec_sw, dec_bne,
dec_andhi, dec_cmpe, dec_cmpg, dec_cmpge, dec_cmpgeu, dec_cmpgu, dec_orhi,
dec_cmpne,
dec_sru, dec_nor, dec_mul, dec_divu, dec_rcsr, dec_sr, dec_xor, dec_ill,
dec_and, dec_xnor, dec_ill, dec_scall, dec_sextb, dec_add, dec_or, dec_sl,
dec_b, dec_modu, dec_sub, dec_user, dec_wcsr, dec_ill, dec_call, dec_sexth,
dec_bi, dec_cmpe, dec_cmpg, dec_cmpge, dec_cmpgeu, dec_cmpgu, dec_calli,
dec_cmpne
};
static inline void decode(DisasContext *dc, uint32_t ir)
{
if (unlikely(qemu_loglevel_mask(CPU_LOG_TB_OP))) {
tcg_gen_debug_insn_start(dc->pc);
}
dc->ir = ir;
LOG_DIS("%8.8x\t", dc->ir);
/* try guessing 'empty' instruction memory, although it may be a valid
* instruction sequence (eg. srui r0, r0, 0) */
if (dc->ir) {
dc->nr_nops = 0;
} else {
LOG_DIS("nr_nops=%d\t", dc->nr_nops);
dc->nr_nops++;
if (dc->nr_nops > 4) {
cpu_abort(dc->env, "fetching nop sequence\n");
}
}
dc->opcode = EXTRACT_FIELD(ir, 26, 31);
dc->imm5 = EXTRACT_FIELD(ir, 0, 4);
dc->imm16 = EXTRACT_FIELD(ir, 0, 15);
dc->imm26 = EXTRACT_FIELD(ir, 0, 25);
dc->csr = EXTRACT_FIELD(ir, 21, 25);
dc->r0 = EXTRACT_FIELD(ir, 21, 25);
dc->r1 = EXTRACT_FIELD(ir, 16, 20);
dc->r2 = EXTRACT_FIELD(ir, 11, 15);
/* bit 31 seems to indicate insn type. */
if (ir & (1 << 31)) {
dc->format = OP_FMT_RR;
} else {
dc->format = OP_FMT_RI;
}
assert(ARRAY_SIZE(decinfo) == 64);
assert(dc->opcode < 64);
decinfo[dc->opcode](dc);
}
static void check_breakpoint(CPULM32State *env, DisasContext *dc)
{
CPUBreakpoint *bp;
if (unlikely(!QTAILQ_EMPTY(&env->breakpoints))) {
QTAILQ_FOREACH(bp, &env->breakpoints, entry) {
if (bp->pc == dc->pc) {
tcg_gen_movi_tl(cpu_pc, dc->pc);
t_gen_raise_exception(dc, EXCP_DEBUG);
dc->is_jmp = DISAS_UPDATE;
}
}
}
}
/* generate intermediate code for basic block 'tb'. */
static void gen_intermediate_code_internal(CPULM32State *env,
TranslationBlock *tb, int search_pc)
{
struct DisasContext ctx, *dc = &ctx;
uint16_t *gen_opc_end;
uint32_t pc_start;
int j, lj;
uint32_t next_page_start;
int num_insns;
int max_insns;
qemu_log_try_set_file(stderr);
pc_start = tb->pc;
dc->env = env;
dc->tb = tb;
gen_opc_end = gen_opc_buf + OPC_MAX_SIZE;
dc->is_jmp = DISAS_NEXT;
dc->pc = pc_start;
dc->singlestep_enabled = env->singlestep_enabled;
dc->nr_nops = 0;
if (pc_start & 3) {
cpu_abort(env, "LM32: unaligned PC=%x\n", pc_start);
}
if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
qemu_log("-----------------------------------------\n");
log_cpu_state(env, 0);
}
next_page_start = (pc_start & TARGET_PAGE_MASK) + TARGET_PAGE_SIZE;
lj = -1;
num_insns = 0;
max_insns = tb->cflags & CF_COUNT_MASK;
if (max_insns == 0) {
max_insns = CF_COUNT_MASK;
}
gen_icount_start();
do {
check_breakpoint(env, dc);
if (search_pc) {
j = gen_opc_ptr - gen_opc_buf;
if (lj < j) {
lj++;
while (lj < j) {
gen_opc_instr_start[lj++] = 0;
}
}
gen_opc_pc[lj] = dc->pc;
gen_opc_instr_start[lj] = 1;
gen_opc_icount[lj] = num_insns;
}
/* Pretty disas. */
LOG_DIS("%8.8x:\t", dc->pc);
if (num_insns + 1 == max_insns && (tb->cflags & CF_LAST_IO)) {
gen_io_start();
}
decode(dc, cpu_ldl_code(env, dc->pc));
dc->pc += 4;
num_insns++;
} while (!dc->is_jmp
&& gen_opc_ptr < gen_opc_end
&& !env->singlestep_enabled
&& !singlestep
&& (dc->pc < next_page_start)
&& num_insns < max_insns);
if (tb->cflags & CF_LAST_IO) {
gen_io_end();
}
if (unlikely(env->singlestep_enabled)) {
if (dc->is_jmp == DISAS_NEXT) {
tcg_gen_movi_tl(cpu_pc, dc->pc);
}
t_gen_raise_exception(dc, EXCP_DEBUG);
} else {
switch (dc->is_jmp) {
case DISAS_NEXT:
gen_goto_tb(dc, 1, dc->pc);
break;
default:
case DISAS_JUMP:
case DISAS_UPDATE:
/* indicate that the hash table must be used
to find the next TB */
tcg_gen_exit_tb(0);
break;
case DISAS_TB_JUMP:
/* nothing more to generate */
break;
}
}
gen_icount_end(tb, num_insns);
*gen_opc_ptr = INDEX_op_end;
if (search_pc) {
j = gen_opc_ptr - gen_opc_buf;
lj++;
while (lj <= j) {
gen_opc_instr_start[lj++] = 0;
}
} else {
tb->size = dc->pc - pc_start;
tb->icount = num_insns;
}
#ifdef DEBUG_DISAS
if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)) {
qemu_log("\n");
log_target_disas(pc_start, dc->pc - pc_start, 0);
qemu_log("\nisize=%d osize=%td\n",
dc->pc - pc_start, gen_opc_ptr - gen_opc_buf);
}
#endif
}
void gen_intermediate_code(CPULM32State *env, struct TranslationBlock *tb)
{
gen_intermediate_code_internal(env, tb, 0);
}
void gen_intermediate_code_pc(CPULM32State *env, struct TranslationBlock *tb)
{
gen_intermediate_code_internal(env, tb, 1);
}
void cpu_dump_state(CPULM32State *env, FILE *f, fprintf_function cpu_fprintf,
int flags)
{
int i;
if (!env || !f) {
return;
}
cpu_fprintf(f, "IN: PC=%x %s\n",
env->pc, lookup_symbol(env->pc));
cpu_fprintf(f, "ie=%8.8x (IE=%x EIE=%x BIE=%x) im=%8.8x ip=%8.8x\n",
env->ie,
(env->ie & IE_IE) ? 1 : 0,
(env->ie & IE_EIE) ? 1 : 0,
(env->ie & IE_BIE) ? 1 : 0,
lm32_pic_get_im(env->pic_state),
lm32_pic_get_ip(env->pic_state));
cpu_fprintf(f, "eba=%8.8x deba=%8.8x\n",
env->eba,
env->deba);
for (i = 0; i < 32; i++) {
cpu_fprintf(f, "r%2.2d=%8.8x ", i, env->regs[i]);
if ((i + 1) % 4 == 0) {
cpu_fprintf(f, "\n");
}
}
cpu_fprintf(f, "\n\n");
}
void restore_state_to_opc(CPULM32State *env, TranslationBlock *tb, int pc_pos)
{
env->pc = gen_opc_pc[pc_pos];
}
void lm32_translate_init(void)
{
int i;
cpu_env = tcg_global_reg_new_ptr(TCG_AREG0, "env");
for (i = 0; i < ARRAY_SIZE(cpu_R); i++) {
cpu_R[i] = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, regs[i]),
regnames[i]);
}
for (i = 0; i < ARRAY_SIZE(cpu_bp); i++) {
cpu_bp[i] = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, bp[i]),
regnames[32+i]);
}
for (i = 0; i < ARRAY_SIZE(cpu_wp); i++) {
cpu_wp[i] = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, wp[i]),
regnames[36+i]);
}
cpu_pc = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, pc),
"pc");
cpu_ie = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, ie),
"ie");
cpu_icc = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, icc),
"icc");
cpu_dcc = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, dcc),
"dcc");
cpu_cc = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, cc),
"cc");
cpu_cfg = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, cfg),
"cfg");
cpu_eba = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, eba),
"eba");
cpu_dc = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, dc),
"dc");
cpu_deba = tcg_global_mem_new(TCG_AREG0,
offsetof(CPULM32State, deba),
"deba");
}