helper.c 41.0 KB
Newer Older
B
bellard 已提交
1
/*
2
 *  i386 helpers (without register variable usage)
3
 *
B
bellard 已提交
4 5 6 7 8 9 10 11 12 13 14 15 16
 *  Copyright (c) 2003 Fabrice Bellard
 *
 * 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
17
 * License along with this library; if not, see <http://www.gnu.org/licenses/>.
B
bellard 已提交
18 19
 */

20
#include "cpu.h"
21
#include "sysemu/kvm.h"
J
Jan Kiszka 已提交
22
#ifndef CONFIG_USER_ONLY
23
#include "sysemu/sysemu.h"
24
#include "monitor/monitor.h"
J
Jan Kiszka 已提交
25
#endif
B
bellard 已提交
26

27
//#define DEBUG_MMU
28

29
static void cpu_x86_version(CPUX86State *env, int *family, int *model)
30 31 32 33 34 35 36 37 38 39 40 41
{
    int cpuver = env->cpuid_version;

    if (family == NULL || model == NULL) {
        return;
    }

    *family = (cpuver >> 8) & 0x0f;
    *model = ((cpuver >> 12) & 0xf0) + ((cpuver >> 4) & 0x0f);
}

/* Broadcast MCA signal for processor version 06H_EH and above */
42
int cpu_x86_support_mca_broadcast(CPUX86State *env)
43 44 45 46 47 48 49 50 51 52 53 54
{
    int family = 0;
    int model = 0;

    cpu_x86_version(env, &family, &model);
    if ((family == 6 && model >= 14) || family > 6) {
        return 1;
    }

    return 0;
}

55 56
/***********************************************************/
/* x86 debug */
57

58
static const char *cc_op_str[CC_OP_NB] = {
59 60
    "DYNAMIC",
    "EFLAGS",
61

62 63 64 65
    "MULB",
    "MULW",
    "MULL",
    "MULQ",
66

67 68 69 70
    "ADDB",
    "ADDW",
    "ADDL",
    "ADDQ",
71

72 73 74 75
    "ADCB",
    "ADCW",
    "ADCL",
    "ADCQ",
76

77 78 79 80
    "SUBB",
    "SUBW",
    "SUBL",
    "SUBQ",
81

82 83 84 85
    "SBBB",
    "SBBW",
    "SBBL",
    "SBBQ",
86

87 88 89 90
    "LOGICB",
    "LOGICW",
    "LOGICL",
    "LOGICQ",
91

92 93 94 95
    "INCB",
    "INCW",
    "INCL",
    "INCQ",
96

97 98 99 100
    "DECB",
    "DECW",
    "DECL",
    "DECQ",
101

102 103 104 105
    "SHLB",
    "SHLW",
    "SHLL",
    "SHLQ",
106

107 108 109 110
    "SARB",
    "SARW",
    "SARL",
    "SARQ",
111 112 113 114 115

    "BMILGB",
    "BMILGW",
    "BMILGL",
    "BMILGQ",
116 117 118 119

    "ADCX",
    "ADOX",
    "ADCOX",
R
Richard Henderson 已提交
120 121

    "CLR",
122
};
123

124
static void
125
cpu_x86_dump_seg_cache(CPUX86State *env, FILE *f, fprintf_function cpu_fprintf,
126 127 128 129 130
                       const char *name, struct SegmentCache *sc)
{
#ifdef TARGET_X86_64
    if (env->hflags & HF_CS64_MASK) {
        cpu_fprintf(f, "%-3s=%04x %016" PRIx64 " %08x %08x", name,
131
                    sc->selector, sc->base, sc->limit, sc->flags & 0x00ffff00);
132 133 134 135
    } else
#endif
    {
        cpu_fprintf(f, "%-3s=%04x %08x %08x %08x", name, sc->selector,
136
                    (uint32_t)sc->base, sc->limit, sc->flags & 0x00ffff00);
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
    }

    if (!(env->hflags & HF_PE_MASK) || !(sc->flags & DESC_P_MASK))
        goto done;

    cpu_fprintf(f, " DPL=%d ", (sc->flags & DESC_DPL_MASK) >> DESC_DPL_SHIFT);
    if (sc->flags & DESC_S_MASK) {
        if (sc->flags & DESC_CS_MASK) {
            cpu_fprintf(f, (sc->flags & DESC_L_MASK) ? "CS64" :
                           ((sc->flags & DESC_B_MASK) ? "CS32" : "CS16"));
            cpu_fprintf(f, " [%c%c", (sc->flags & DESC_C_MASK) ? 'C' : '-',
                        (sc->flags & DESC_R_MASK) ? 'R' : '-');
        } else {
            cpu_fprintf(f, (sc->flags & DESC_B_MASK) ? "DS  " : "DS16");
            cpu_fprintf(f, " [%c%c", (sc->flags & DESC_E_MASK) ? 'E' : '-',
                        (sc->flags & DESC_W_MASK) ? 'W' : '-');
        }
        cpu_fprintf(f, "%c]", (sc->flags & DESC_A_MASK) ? 'A' : '-');
    } else {
        static const char *sys_type_name[2][16] = {
            { /* 32 bit mode */
                "Reserved", "TSS16-avl", "LDT", "TSS16-busy",
                "CallGate16", "TaskGate", "IntGate16", "TrapGate16",
                "Reserved", "TSS32-avl", "Reserved", "TSS32-busy",
                "CallGate32", "Reserved", "IntGate32", "TrapGate32"
            },
            { /* 64 bit mode */
                "<hiword>", "Reserved", "LDT", "Reserved", "Reserved",
                "Reserved", "Reserved", "Reserved", "Reserved",
                "TSS64-avl", "Reserved", "TSS64-busy", "CallGate64",
                "Reserved", "IntGate64", "TrapGate64"
            }
        };
S
Stefan Weil 已提交
170 171 172 173
        cpu_fprintf(f, "%s",
                    sys_type_name[(env->hflags & HF_LMA_MASK) ? 1 : 0]
                                 [(sc->flags & DESC_TYPE_MASK)
                                  >> DESC_TYPE_SHIFT]);
174 175 176 177 178
    }
done:
    cpu_fprintf(f, "\n");
}

179 180 181
#define DUMP_CODE_BYTES_TOTAL    50
#define DUMP_CODE_BYTES_BACKWARD 20

182 183
void x86_cpu_dump_state(CPUState *cs, FILE *f, fprintf_function cpu_fprintf,
                        int flags)
184
{
185 186
    X86CPU *cpu = X86_CPU(cs);
    CPUX86State *env = &cpu->env;
187 188 189
    int eflags, i, nb;
    char cc_op_name[32];
    static const char *seg_name[6] = { "ES", "CS", "SS", "DS", "FS", "GS" };
190

191
    cpu_synchronize_state(cs);
B
balrog 已提交
192

193
    eflags = cpu_compute_eflags(env);
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
#ifdef TARGET_X86_64
    if (env->hflags & HF_CS64_MASK) {
        cpu_fprintf(f,
                    "RAX=%016" PRIx64 " RBX=%016" PRIx64 " RCX=%016" PRIx64 " RDX=%016" PRIx64 "\n"
                    "RSI=%016" PRIx64 " RDI=%016" PRIx64 " RBP=%016" PRIx64 " RSP=%016" PRIx64 "\n"
                    "R8 =%016" PRIx64 " R9 =%016" PRIx64 " R10=%016" PRIx64 " R11=%016" PRIx64 "\n"
                    "R12=%016" PRIx64 " R13=%016" PRIx64 " R14=%016" PRIx64 " R15=%016" PRIx64 "\n"
                    "RIP=%016" PRIx64 " RFL=%08x [%c%c%c%c%c%c%c] CPL=%d II=%d A20=%d SMM=%d HLT=%d\n",
                    env->regs[R_EAX],
                    env->regs[R_EBX],
                    env->regs[R_ECX],
                    env->regs[R_EDX],
                    env->regs[R_ESI],
                    env->regs[R_EDI],
                    env->regs[R_EBP],
                    env->regs[R_ESP],
                    env->regs[8],
                    env->regs[9],
                    env->regs[10],
                    env->regs[11],
                    env->regs[12],
                    env->regs[13],
                    env->regs[14],
                    env->regs[15],
                    env->eip, eflags,
                    eflags & DF_MASK ? 'D' : '-',
                    eflags & CC_O ? 'O' : '-',
                    eflags & CC_S ? 'S' : '-',
                    eflags & CC_Z ? 'Z' : '-',
                    eflags & CC_A ? 'A' : '-',
                    eflags & CC_P ? 'P' : '-',
                    eflags & CC_C ? 'C' : '-',
                    env->hflags & HF_CPL_MASK,
                    (env->hflags >> HF_INHIBIT_IRQ_SHIFT) & 1,
J
Juan Quintela 已提交
228
                    (env->a20_mask >> 20) & 1,
229
                    (env->hflags >> HF_SMM_SHIFT) & 1,
230
                    cs->halted);
231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
    } else
#endif
    {
        cpu_fprintf(f, "EAX=%08x EBX=%08x ECX=%08x EDX=%08x\n"
                    "ESI=%08x EDI=%08x EBP=%08x ESP=%08x\n"
                    "EIP=%08x EFL=%08x [%c%c%c%c%c%c%c] CPL=%d II=%d A20=%d SMM=%d HLT=%d\n",
                    (uint32_t)env->regs[R_EAX],
                    (uint32_t)env->regs[R_EBX],
                    (uint32_t)env->regs[R_ECX],
                    (uint32_t)env->regs[R_EDX],
                    (uint32_t)env->regs[R_ESI],
                    (uint32_t)env->regs[R_EDI],
                    (uint32_t)env->regs[R_EBP],
                    (uint32_t)env->regs[R_ESP],
                    (uint32_t)env->eip, eflags,
                    eflags & DF_MASK ? 'D' : '-',
                    eflags & CC_O ? 'O' : '-',
                    eflags & CC_S ? 'S' : '-',
                    eflags & CC_Z ? 'Z' : '-',
                    eflags & CC_A ? 'A' : '-',
                    eflags & CC_P ? 'P' : '-',
                    eflags & CC_C ? 'C' : '-',
                    env->hflags & HF_CPL_MASK,
                    (env->hflags >> HF_INHIBIT_IRQ_SHIFT) & 1,
J
Juan Quintela 已提交
255
                    (env->a20_mask >> 20) & 1,
256
                    (env->hflags >> HF_SMM_SHIFT) & 1,
257
                    cs->halted);
258
    }
259

260 261 262 263 264 265 266
    for(i = 0; i < 6; i++) {
        cpu_x86_dump_seg_cache(env, f, cpu_fprintf, seg_name[i],
                               &env->segs[i]);
    }
    cpu_x86_dump_seg_cache(env, f, cpu_fprintf, "LDT", &env->ldt);
    cpu_x86_dump_seg_cache(env, f, cpu_fprintf, "TR", &env->tr);

267 268 269 270 271 272 273 274 275 276 277
#ifdef TARGET_X86_64
    if (env->hflags & HF_LMA_MASK) {
        cpu_fprintf(f, "GDT=     %016" PRIx64 " %08x\n",
                    env->gdt.base, env->gdt.limit);
        cpu_fprintf(f, "IDT=     %016" PRIx64 " %08x\n",
                    env->idt.base, env->idt.limit);
        cpu_fprintf(f, "CR0=%08x CR2=%016" PRIx64 " CR3=%016" PRIx64 " CR4=%08x\n",
                    (uint32_t)env->cr[0],
                    env->cr[2],
                    env->cr[3],
                    (uint32_t)env->cr[4]);
A
aliguori 已提交
278 279 280
        for(i = 0; i < 4; i++)
            cpu_fprintf(f, "DR%d=%016" PRIx64 " ", i, env->dr[i]);
        cpu_fprintf(f, "\nDR6=%016" PRIx64 " DR7=%016" PRIx64 "\n",
281
                    env->dr[6], env->dr[7]);
282 283 284 285 286 287 288 289 290 291 292 293
    } else
#endif
    {
        cpu_fprintf(f, "GDT=     %08x %08x\n",
                    (uint32_t)env->gdt.base, env->gdt.limit);
        cpu_fprintf(f, "IDT=     %08x %08x\n",
                    (uint32_t)env->idt.base, env->idt.limit);
        cpu_fprintf(f, "CR0=%08x CR2=%08x CR3=%08x CR4=%08x\n",
                    (uint32_t)env->cr[0],
                    (uint32_t)env->cr[2],
                    (uint32_t)env->cr[3],
                    (uint32_t)env->cr[4]);
294 295 296 297 298
        for(i = 0; i < 4; i++) {
            cpu_fprintf(f, "DR%d=" TARGET_FMT_lx " ", i, env->dr[i]);
        }
        cpu_fprintf(f, "\nDR6=" TARGET_FMT_lx " DR7=" TARGET_FMT_lx "\n",
                    env->dr[6], env->dr[7]);
299
    }
300
    if (flags & CPU_DUMP_CCOP) {
301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
        if ((unsigned)env->cc_op < CC_OP_NB)
            snprintf(cc_op_name, sizeof(cc_op_name), "%s", cc_op_str[env->cc_op]);
        else
            snprintf(cc_op_name, sizeof(cc_op_name), "[%d]", env->cc_op);
#ifdef TARGET_X86_64
        if (env->hflags & HF_CS64_MASK) {
            cpu_fprintf(f, "CCS=%016" PRIx64 " CCD=%016" PRIx64 " CCO=%-8s\n",
                        env->cc_src, env->cc_dst,
                        cc_op_name);
        } else
#endif
        {
            cpu_fprintf(f, "CCS=%08x CCD=%08x CCO=%-8s\n",
                        (uint32_t)env->cc_src, (uint32_t)env->cc_dst,
                        cc_op_name);
        }
317
    }
318
    cpu_fprintf(f, "EFER=%016" PRIx64 "\n", env->efer);
319
    if (flags & CPU_DUMP_FPU) {
320 321 322 323 324 325 326 327 328 329 330 331
        int fptag;
        fptag = 0;
        for(i = 0; i < 8; i++) {
            fptag |= ((!env->fptags[i]) << i);
        }
        cpu_fprintf(f, "FCW=%04x FSW=%04x [ST=%d] FTW=%02x MXCSR=%08x\n",
                    env->fpuc,
                    (env->fpus & ~0x3800) | (env->fpstt & 0x7) << 11,
                    env->fpstt,
                    fptag,
                    env->mxcsr);
        for(i=0;i<8;i++) {
332 333
            CPU_LDoubleU u;
            u.d = env->fpregs[i].d;
334
            cpu_fprintf(f, "FPR%d=%016" PRIx64 " %04x",
335
                        i, u.l.lower, u.l.upper);
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
            if ((i & 1) == 1)
                cpu_fprintf(f, "\n");
            else
                cpu_fprintf(f, " ");
        }
        if (env->hflags & HF_CS64_MASK)
            nb = 16;
        else
            nb = 8;
        for(i=0;i<nb;i++) {
            cpu_fprintf(f, "XMM%02d=%08x%08x%08x%08x",
                        i,
                        env->xmm_regs[i].XMM_L(3),
                        env->xmm_regs[i].XMM_L(2),
                        env->xmm_regs[i].XMM_L(1),
                        env->xmm_regs[i].XMM_L(0));
            if ((i & 1) == 1)
                cpu_fprintf(f, "\n");
            else
                cpu_fprintf(f, " ");
        }
357
    }
358 359 360 361 362 363 364 365
    if (flags & CPU_DUMP_CODE) {
        target_ulong base = env->segs[R_CS].base + env->eip;
        target_ulong offs = MIN(env->eip, DUMP_CODE_BYTES_BACKWARD);
        uint8_t code;
        char codestr[3];

        cpu_fprintf(f, "Code=");
        for (i = 0; i < DUMP_CODE_BYTES_TOTAL; i++) {
366
            if (cpu_memory_rw_debug(cs, base - offs + i, &code, 1, 0) == 0) {
367 368 369 370 371 372 373 374 375
                snprintf(codestr, sizeof(codestr), "%02x", code);
            } else {
                snprintf(codestr, sizeof(codestr), "??");
            }
            cpu_fprintf(f, "%s%s%s%s", i > 0 ? " " : "",
                        i == offs ? "<" : "", codestr, i == offs ? ">" : "");
        }
        cpu_fprintf(f, "\n");
    }
B
bellard 已提交
376
}
377

378 379 380 381
/***********************************************************/
/* x86 mmu */
/* XXX: add PGE support */

382
void x86_cpu_set_a20(X86CPU *cpu, int a20_state)
B
bellard 已提交
383
{
384 385
    CPUX86State *env = &cpu->env;

386 387 388 389 390 391 392
    a20_state = (a20_state != 0);
    if (a20_state != ((env->a20_mask >> 20) & 1)) {
#if defined(DEBUG_MMU)
        printf("A20 update: a20=%d\n", a20_state);
#endif
        /* if the cpu is currently executing code, we must unlink it and
           all the potentially executing TB */
393
        cpu_interrupt(CPU(cpu), CPU_INTERRUPT_EXITTB);
394

395 396 397
        /* when a20 is changed, all the MMU mappings are invalid, so
           we must flush everything */
        tlb_flush(env, 1);
J
Juan Quintela 已提交
398
        env->a20_mask = ~(1 << 20) | (a20_state << 20);
399
    }
B
bellard 已提交
400 401
}

402
void cpu_x86_update_cr0(CPUX86State *env, uint32_t new_cr0)
B
bellard 已提交
403
{
404
    int pe_state;
B
bellard 已提交
405

406 407 408 409 410 411 412
#if defined(DEBUG_MMU)
    printf("CR0 update: CR0=0x%08x\n", new_cr0);
#endif
    if ((new_cr0 & (CR0_PG_MASK | CR0_WP_MASK | CR0_PE_MASK)) !=
        (env->cr[0] & (CR0_PG_MASK | CR0_WP_MASK | CR0_PE_MASK))) {
        tlb_flush(env, 1);
    }
B
bellard 已提交
413

414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431
#ifdef TARGET_X86_64
    if (!(env->cr[0] & CR0_PG_MASK) && (new_cr0 & CR0_PG_MASK) &&
        (env->efer & MSR_EFER_LME)) {
        /* enter in long mode */
        /* XXX: generate an exception */
        if (!(env->cr[4] & CR4_PAE_MASK))
            return;
        env->efer |= MSR_EFER_LMA;
        env->hflags |= HF_LMA_MASK;
    } else if ((env->cr[0] & CR0_PG_MASK) && !(new_cr0 & CR0_PG_MASK) &&
               (env->efer & MSR_EFER_LMA)) {
        /* exit long mode */
        env->efer &= ~MSR_EFER_LMA;
        env->hflags &= ~(HF_LMA_MASK | HF_CS64_MASK);
        env->eip &= 0xffffffff;
    }
#endif
    env->cr[0] = new_cr0 | CR0_ET_MASK;
432

433 434 435 436 437 438 439 440
    /* update PE flag in hidden flags */
    pe_state = (env->cr[0] & CR0_PE_MASK);
    env->hflags = (env->hflags & ~HF_PE_MASK) | (pe_state << HF_PE_SHIFT);
    /* ensure that ADDSEG is always set in real mode */
    env->hflags |= ((pe_state ^ 1) << HF_ADDSEG_SHIFT);
    /* update FPU flags */
    env->hflags = (env->hflags & ~(HF_MP_MASK | HF_EM_MASK | HF_TS_MASK)) |
        ((new_cr0 << (HF_MP_SHIFT - 1)) & (HF_MP_MASK | HF_EM_MASK | HF_TS_MASK));
441 442
}

443 444 445
/* XXX: in legacy PAE mode, generate a GPF if reserved bits are set in
   the PDPT */
void cpu_x86_update_cr3(CPUX86State *env, target_ulong new_cr3)
446
{
447 448 449 450 451 452 453
    env->cr[3] = new_cr3;
    if (env->cr[0] & CR0_PG_MASK) {
#if defined(DEBUG_MMU)
        printf("CR3 update: CR3=" TARGET_FMT_lx "\n", new_cr3);
#endif
        tlb_flush(env, 0);
    }
454 455
}

456
void cpu_x86_update_cr4(CPUX86State *env, uint32_t new_cr4)
457
{
458 459 460
#if defined(DEBUG_MMU)
    printf("CR4 update: CR4=%08x\n", (uint32_t)env->cr[4]);
#endif
H
H. Peter Anvin 已提交
461 462 463
    if ((new_cr4 ^ env->cr[4]) &
        (CR4_PGE_MASK | CR4_PAE_MASK | CR4_PSE_MASK |
         CR4_SMEP_MASK | CR4_SMAP_MASK)) {
464 465 466
        tlb_flush(env, 1);
    }
    /* SSE handling */
467
    if (!(env->features[FEAT_1_EDX] & CPUID_SSE)) {
468
        new_cr4 &= ~CR4_OSFXSR_MASK;
H
H. Peter Anvin 已提交
469 470 471
    }
    env->hflags &= ~HF_OSFXSR_MASK;
    if (new_cr4 & CR4_OSFXSR_MASK) {
472
        env->hflags |= HF_OSFXSR_MASK;
H
H. Peter Anvin 已提交
473 474
    }

475
    if (!(env->features[FEAT_7_0_EBX] & CPUID_7_0_EBX_SMAP)) {
H
H. Peter Anvin 已提交
476 477 478 479 480 481
        new_cr4 &= ~CR4_SMAP_MASK;
    }
    env->hflags &= ~HF_SMAP_MASK;
    if (new_cr4 & CR4_SMAP_MASK) {
        env->hflags |= HF_SMAP_MASK;
    }
482

483
    env->cr[4] = new_cr4;
484 485
}

486 487 488
#if defined(CONFIG_USER_ONLY)

int cpu_x86_handle_mmu_fault(CPUX86State *env, target_ulong addr,
489
                             int is_write, int mmu_idx)
490
{
491 492 493 494 495 496 497
    /* user mode only emulation */
    is_write &= 1;
    env->cr[2] = addr;
    env->error_code = (is_write << PG_ERROR_W_BIT);
    env->error_code |= PG_ERROR_U_MASK;
    env->exception_index = EXCP0E_PAGE;
    return 1;
B
bellard 已提交
498 499
}

500
#else
501

502 503 504
/* XXX: This value should match the one returned by CPUID
 * and in exec.c */
# if defined(TARGET_X86_64)
505
# define PHYS_ADDR_MASK 0xfffffff000LL
506
# else
507
# define PHYS_ADDR_MASK 0xffffff000LL
508 509 510 511 512 513 514 515
# endif

/* return value:
   -1 = cannot handle fault
   0  = nothing more to do
   1  = generate PF fault
*/
int cpu_x86_handle_mmu_fault(CPUX86State *env, target_ulong addr,
516
                             int is_write1, int mmu_idx)
517 518 519
{
    uint64_t ptep, pte;
    target_ulong pde_addr, pte_addr;
P
Paul Brook 已提交
520
    int error_code, is_dirty, prot, page_size, is_write, is_user;
A
Avi Kivity 已提交
521
    hwaddr paddr;
522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542
    uint32_t page_offset;
    target_ulong vaddr, virt_addr;

    is_user = mmu_idx == MMU_USER_IDX;
#if defined(DEBUG_MMU)
    printf("MMU fault: addr=" TARGET_FMT_lx " w=%d u=%d eip=" TARGET_FMT_lx "\n",
           addr, is_write1, is_user, env->eip);
#endif
    is_write = is_write1 & 1;

    if (!(env->cr[0] & CR0_PG_MASK)) {
        pte = addr;
        virt_addr = addr & TARGET_PAGE_MASK;
        prot = PAGE_READ | PAGE_WRITE | PAGE_EXEC;
        page_size = 4096;
        goto do_mapping;
    }

    if (env->cr[4] & CR4_PAE_MASK) {
        uint64_t pde, pdpe;
        target_ulong pdpe_addr;
B
bellard 已提交
543

544 545 546 547 548 549 550 551 552 553 554 555
#ifdef TARGET_X86_64
        if (env->hflags & HF_LMA_MASK) {
            uint64_t pml4e_addr, pml4e;
            int32_t sext;

            /* test virtual address sign extension */
            sext = (int64_t)addr >> 47;
            if (sext != 0 && sext != -1) {
                env->error_code = 0;
                env->exception_index = EXCP0D_GPF;
                return 1;
            }
T
ths 已提交
556

557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
            pml4e_addr = ((env->cr[3] & ~0xfff) + (((addr >> 39) & 0x1ff) << 3)) &
                env->a20_mask;
            pml4e = ldq_phys(pml4e_addr);
            if (!(pml4e & PG_PRESENT_MASK)) {
                error_code = 0;
                goto do_fault;
            }
            if (!(env->efer & MSR_EFER_NXE) && (pml4e & PG_NX_MASK)) {
                error_code = PG_ERROR_RSVD_MASK;
                goto do_fault;
            }
            if (!(pml4e & PG_ACCESSED_MASK)) {
                pml4e |= PG_ACCESSED_MASK;
                stl_phys_notdirty(pml4e_addr, pml4e);
            }
            ptep = pml4e ^ PG_NX_MASK;
            pdpe_addr = ((pml4e & PHYS_ADDR_MASK) + (((addr >> 30) & 0x1ff) << 3)) &
                env->a20_mask;
            pdpe = ldq_phys(pdpe_addr);
            if (!(pdpe & PG_PRESENT_MASK)) {
                error_code = 0;
                goto do_fault;
            }
            if (!(env->efer & MSR_EFER_NXE) && (pdpe & PG_NX_MASK)) {
                error_code = PG_ERROR_RSVD_MASK;
                goto do_fault;
            }
            ptep &= pdpe ^ PG_NX_MASK;
            if (!(pdpe & PG_ACCESSED_MASK)) {
                pdpe |= PG_ACCESSED_MASK;
                stl_phys_notdirty(pdpe_addr, pdpe);
            }
        } else
#endif
        {
            /* XXX: load them when cr3 is loaded ? */
            pdpe_addr = ((env->cr[3] & ~0x1f) + ((addr >> 27) & 0x18)) &
                env->a20_mask;
            pdpe = ldq_phys(pdpe_addr);
            if (!(pdpe & PG_PRESENT_MASK)) {
                error_code = 0;
                goto do_fault;
            }
            ptep = PG_NX_MASK | PG_USER_MASK | PG_RW_MASK;
601 602
        }

603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618
        pde_addr = ((pdpe & PHYS_ADDR_MASK) + (((addr >> 21) & 0x1ff) << 3)) &
            env->a20_mask;
        pde = ldq_phys(pde_addr);
        if (!(pde & PG_PRESENT_MASK)) {
            error_code = 0;
            goto do_fault;
        }
        if (!(env->efer & MSR_EFER_NXE) && (pde & PG_NX_MASK)) {
            error_code = PG_ERROR_RSVD_MASK;
            goto do_fault;
        }
        ptep &= pde ^ PG_NX_MASK;
        if (pde & PG_PSE_MASK) {
            /* 2 MB page */
            page_size = 2048 * 1024;
            ptep ^= PG_NX_MASK;
H
H. Peter Anvin 已提交
619
            if ((ptep & PG_NX_MASK) && is_write1 == 2) {
620
                goto do_fault_protect;
H
H. Peter Anvin 已提交
621 622 623 624
            }
            switch (mmu_idx) {
            case MMU_USER_IDX:
                if (!(ptep & PG_USER_MASK)) {
625
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
626 627
                }
                if (is_write && !(ptep & PG_RW_MASK)) {
628
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
629 630 631 632 633 634 635 636 637 638 639 640 641 642
                }
                break;

            case MMU_KERNEL_IDX:
                if (is_write1 != 2 && (env->cr[4] & CR4_SMAP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
                /* fall through */
            case MMU_KSMAP_IDX:
                if (is_write1 == 2 && (env->cr[4] & CR4_SMEP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
643
                if ((env->cr[0] & CR0_WP_MASK) &&
H
H. Peter Anvin 已提交
644
                    is_write && !(ptep & PG_RW_MASK)) {
645
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
646 647 648 649 650
                }
                break;

            default: /* cannot happen */
                break;
651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
            }
            is_dirty = is_write && !(pde & PG_DIRTY_MASK);
            if (!(pde & PG_ACCESSED_MASK) || is_dirty) {
                pde |= PG_ACCESSED_MASK;
                if (is_dirty)
                    pde |= PG_DIRTY_MASK;
                stl_phys_notdirty(pde_addr, pde);
            }
            /* align to page_size */
            pte = pde & ((PHYS_ADDR_MASK & ~(page_size - 1)) | 0xfff);
            virt_addr = addr & ~(page_size - 1);
        } else {
            /* 4 KB page */
            if (!(pde & PG_ACCESSED_MASK)) {
                pde |= PG_ACCESSED_MASK;
                stl_phys_notdirty(pde_addr, pde);
            }
            pte_addr = ((pde & PHYS_ADDR_MASK) + (((addr >> 12) & 0x1ff) << 3)) &
                env->a20_mask;
            pte = ldq_phys(pte_addr);
            if (!(pte & PG_PRESENT_MASK)) {
                error_code = 0;
                goto do_fault;
            }
            if (!(env->efer & MSR_EFER_NXE) && (pte & PG_NX_MASK)) {
                error_code = PG_ERROR_RSVD_MASK;
                goto do_fault;
            }
            /* combine pde and pte nx, user and rw protections */
            ptep &= pte ^ PG_NX_MASK;
            ptep ^= PG_NX_MASK;
            if ((ptep & PG_NX_MASK) && is_write1 == 2)
                goto do_fault_protect;
H
H. Peter Anvin 已提交
684 685 686
            switch (mmu_idx) {
            case MMU_USER_IDX:
                if (!(ptep & PG_USER_MASK)) {
687
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
688 689
                }
                if (is_write && !(ptep & PG_RW_MASK)) {
690
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
691 692 693 694 695 696 697 698 699 700 701 702 703 704
                }
                break;

            case MMU_KERNEL_IDX:
                if (is_write1 != 2 && (env->cr[4] & CR4_SMAP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
                /* fall through */
            case MMU_KSMAP_IDX:
                if (is_write1 == 2 && (env->cr[4] & CR4_SMEP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
705
                if ((env->cr[0] & CR0_WP_MASK) &&
H
H. Peter Anvin 已提交
706
                    is_write && !(ptep & PG_RW_MASK)) {
707
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
708 709 710 711 712
                }
                break;

            default: /* cannot happen */
                break;
713 714 715 716 717 718 719 720 721 722 723
            }
            is_dirty = is_write && !(pte & PG_DIRTY_MASK);
            if (!(pte & PG_ACCESSED_MASK) || is_dirty) {
                pte |= PG_ACCESSED_MASK;
                if (is_dirty)
                    pte |= PG_DIRTY_MASK;
                stl_phys_notdirty(pte_addr, pte);
            }
            page_size = 4096;
            virt_addr = addr & ~0xfff;
            pte = pte & (PHYS_ADDR_MASK | 0xfff);
724
        }
B
bellard 已提交
725
    } else {
726 727 728 729 730 731 732 733 734 735 736 737 738
        uint32_t pde;

        /* page directory entry */
        pde_addr = ((env->cr[3] & ~0xfff) + ((addr >> 20) & 0xffc)) &
            env->a20_mask;
        pde = ldl_phys(pde_addr);
        if (!(pde & PG_PRESENT_MASK)) {
            error_code = 0;
            goto do_fault;
        }
        /* if PSE bit is set, then we use a 4MB page */
        if ((pde & PG_PSE_MASK) && (env->cr[4] & CR4_PSE_MASK)) {
            page_size = 4096 * 1024;
H
H. Peter Anvin 已提交
739 740 741
            switch (mmu_idx) {
            case MMU_USER_IDX:
                if (!(pde & PG_USER_MASK)) {
742
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
743 744
                }
                if (is_write && !(pde & PG_RW_MASK)) {
745
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
746 747 748 749 750 751 752 753 754 755 756 757 758 759
                }
                break;

            case MMU_KERNEL_IDX:
                if (is_write1 != 2 && (env->cr[4] & CR4_SMAP_MASK) &&
                    (pde & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
                /* fall through */
            case MMU_KSMAP_IDX:
                if (is_write1 == 2 && (env->cr[4] & CR4_SMEP_MASK) &&
                    (pde & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
760
                if ((env->cr[0] & CR0_WP_MASK) &&
H
H. Peter Anvin 已提交
761
                    is_write && !(pde & PG_RW_MASK)) {
762
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
763 764 765 766 767
                }
                break;

            default: /* cannot happen */
                break;
768 769 770 771 772 773 774 775
            }
            is_dirty = is_write && !(pde & PG_DIRTY_MASK);
            if (!(pde & PG_ACCESSED_MASK) || is_dirty) {
                pde |= PG_ACCESSED_MASK;
                if (is_dirty)
                    pde |= PG_DIRTY_MASK;
                stl_phys_notdirty(pde_addr, pde);
            }
B
bellard 已提交
776

777 778 779 780 781 782 783 784
            pte = pde & ~( (page_size - 1) & ~0xfff); /* align to page_size */
            ptep = pte;
            virt_addr = addr & ~(page_size - 1);
        } else {
            if (!(pde & PG_ACCESSED_MASK)) {
                pde |= PG_ACCESSED_MASK;
                stl_phys_notdirty(pde_addr, pde);
            }
785

786 787 788 789 790 791 792
            /* page directory entry */
            pte_addr = ((pde & ~0xfff) + ((addr >> 10) & 0xffc)) &
                env->a20_mask;
            pte = ldl_phys(pte_addr);
            if (!(pte & PG_PRESENT_MASK)) {
                error_code = 0;
                goto do_fault;
793
            }
794 795
            /* combine pde and pte user and rw protections */
            ptep = pte & pde;
H
H. Peter Anvin 已提交
796 797 798
            switch (mmu_idx) {
            case MMU_USER_IDX:
                if (!(ptep & PG_USER_MASK)) {
799
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
800 801
                }
                if (is_write && !(ptep & PG_RW_MASK)) {
802
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
803 804 805 806 807 808 809 810 811 812 813 814 815 816
                }
                break;

            case MMU_KERNEL_IDX:
                if (is_write1 != 2 && (env->cr[4] & CR4_SMAP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
                /* fall through */
            case MMU_KSMAP_IDX:
                if (is_write1 == 2 && (env->cr[4] & CR4_SMEP_MASK) &&
                    (ptep & PG_USER_MASK)) {
                    goto do_fault_protect;
                }
817
                if ((env->cr[0] & CR0_WP_MASK) &&
H
H. Peter Anvin 已提交
818
                    is_write && !(ptep & PG_RW_MASK)) {
819
                    goto do_fault_protect;
H
H. Peter Anvin 已提交
820 821 822 823 824
                }
                break;

            default: /* cannot happen */
                break;
825
            }
826 827 828 829 830 831 832 833 834
            is_dirty = is_write && !(pte & PG_DIRTY_MASK);
            if (!(pte & PG_ACCESSED_MASK) || is_dirty) {
                pte |= PG_ACCESSED_MASK;
                if (is_dirty)
                    pte |= PG_DIRTY_MASK;
                stl_phys_notdirty(pte_addr, pte);
            }
            page_size = 4096;
            virt_addr = addr & ~0xfff;
B
bellard 已提交
835 836
        }
    }
837 838 839 840 841 842 843 844 845 846 847 848 849 850
    /* the page can be put in the TLB */
    prot = PAGE_READ;
    if (!(ptep & PG_NX_MASK))
        prot |= PAGE_EXEC;
    if (pte & PG_DIRTY_MASK) {
        /* only set write access if already dirty... otherwise wait
           for dirty access */
        if (is_user) {
            if (ptep & PG_RW_MASK)
                prot |= PAGE_WRITE;
        } else {
            if (!(env->cr[0] & CR0_WP_MASK) ||
                (ptep & PG_RW_MASK))
                prot |= PAGE_WRITE;
851
        }
852
    }
853 854 855 856 857 858 859 860 861
 do_mapping:
    pte = pte & env->a20_mask;

    /* Even if 4MB pages, we map only one 4KB page in the cache to
       avoid filling it too fast */
    page_offset = (addr & TARGET_PAGE_MASK) & (page_size - 1);
    paddr = (pte & TARGET_PAGE_MASK) + page_offset;
    vaddr = virt_addr + page_offset;

P
Paul Brook 已提交
862 863
    tlb_set_page(env, vaddr, paddr, prot, mmu_idx, page_size);
    return 0;
864 865 866 867 868 869 870
 do_fault_protect:
    error_code = PG_ERROR_P_MASK;
 do_fault:
    error_code |= (is_write << PG_ERROR_W_BIT);
    if (is_user)
        error_code |= PG_ERROR_U_MASK;
    if (is_write1 == 2 &&
H
H. Peter Anvin 已提交
871 872 873
        (((env->efer & MSR_EFER_NXE) &&
          (env->cr[4] & CR4_PAE_MASK)) ||
         (env->cr[4] & CR4_SMEP_MASK)))
874
        error_code |= PG_ERROR_I_D_MASK;
B
bellard 已提交
875 876 877 878
    if (env->intercept_exceptions & (1 << EXCP0E_PAGE)) {
        /* cr2 is not modified in case of exceptions */
        stq_phys(env->vm_vmcb + offsetof(struct vmcb, control.exit_info_2), 
                 addr);
879 880
    } else {
        env->cr[2] = addr;
B
bellard 已提交
881
    }
882 883 884
    env->error_code = error_code;
    env->exception_index = EXCP0E_PAGE;
    return 1;
B
bellard 已提交
885 886
}

887
hwaddr x86_cpu_get_phys_page_debug(CPUState *cs, vaddr addr)
B
bellard 已提交
888
{
889 890
    X86CPU *cpu = X86_CPU(cs);
    CPUX86State *env = &cpu->env;
891 892
    target_ulong pde_addr, pte_addr;
    uint64_t pte;
A
Avi Kivity 已提交
893
    hwaddr paddr;
894 895
    uint32_t page_offset;
    int page_size;
B
bellard 已提交
896

897 898 899 900
    if (!(env->cr[0] & CR0_PG_MASK)) {
        pte = addr & env->a20_mask;
        page_size = 4096;
    } else if (env->cr[4] & CR4_PAE_MASK) {
901 902
        target_ulong pdpe_addr;
        uint64_t pde, pdpe;
B
bellard 已提交
903

904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919
#ifdef TARGET_X86_64
        if (env->hflags & HF_LMA_MASK) {
            uint64_t pml4e_addr, pml4e;
            int32_t sext;

            /* test virtual address sign extension */
            sext = (int64_t)addr >> 47;
            if (sext != 0 && sext != -1)
                return -1;

            pml4e_addr = ((env->cr[3] & ~0xfff) + (((addr >> 39) & 0x1ff) << 3)) &
                env->a20_mask;
            pml4e = ldq_phys(pml4e_addr);
            if (!(pml4e & PG_PRESENT_MASK))
                return -1;

920 921
            pdpe_addr = ((pml4e & ~0xfff & ~(PG_NX_MASK | PG_HI_USER_MASK)) +
                         (((addr >> 30) & 0x1ff) << 3)) & env->a20_mask;
922 923 924 925 926 927 928 929 930 931 932
            pdpe = ldq_phys(pdpe_addr);
            if (!(pdpe & PG_PRESENT_MASK))
                return -1;
        } else
#endif
        {
            pdpe_addr = ((env->cr[3] & ~0x1f) + ((addr >> 27) & 0x18)) &
                env->a20_mask;
            pdpe = ldq_phys(pdpe_addr);
            if (!(pdpe & PG_PRESENT_MASK))
                return -1;
B
bellard 已提交
933 934
        }

935 936
        pde_addr = ((pdpe & ~0xfff & ~(PG_NX_MASK | PG_HI_USER_MASK)) +
                    (((addr >> 21) & 0x1ff) << 3)) & env->a20_mask;
937 938 939 940 941 942 943 944 945 946
        pde = ldq_phys(pde_addr);
        if (!(pde & PG_PRESENT_MASK)) {
            return -1;
        }
        if (pde & PG_PSE_MASK) {
            /* 2 MB page */
            page_size = 2048 * 1024;
            pte = pde & ~( (page_size - 1) & ~0xfff); /* align to page_size */
        } else {
            /* 4 KB page */
947 948
            pte_addr = ((pde & ~0xfff & ~(PG_NX_MASK | PG_HI_USER_MASK)) +
                        (((addr >> 12) & 0x1ff) << 3)) & env->a20_mask;
949 950 951
            page_size = 4096;
            pte = ldq_phys(pte_addr);
        }
952
        pte &= ~(PG_NX_MASK | PG_HI_USER_MASK);
953 954
        if (!(pte & PG_PRESENT_MASK))
            return -1;
B
bellard 已提交
955
    } else {
956
        uint32_t pde;
957

958 959 960 961 962 963 964 965
        /* page directory entry */
        pde_addr = ((env->cr[3] & ~0xfff) + ((addr >> 20) & 0xffc)) & env->a20_mask;
        pde = ldl_phys(pde_addr);
        if (!(pde & PG_PRESENT_MASK))
            return -1;
        if ((pde & PG_PSE_MASK) && (env->cr[4] & CR4_PSE_MASK)) {
            pte = pde & ~0x003ff000; /* align to 4MB */
            page_size = 4096 * 1024;
966 967
        } else {
            /* page directory entry */
968 969 970
            pte_addr = ((pde & ~0xfff) + ((addr >> 10) & 0xffc)) & env->a20_mask;
            pte = ldl_phys(pte_addr);
            if (!(pte & PG_PRESENT_MASK))
971
                return -1;
972
            page_size = 4096;
973 974
        }
        pte = pte & env->a20_mask;
B
bellard 已提交
975 976
    }

977 978 979
    page_offset = (addr & TARGET_PAGE_MASK) & (page_size - 1);
    paddr = (pte & TARGET_PAGE_MASK) + page_offset;
    return paddr;
B
bellard 已提交
980
}
981

982
void hw_breakpoint_insert(CPUX86State *env, int index)
983
{
984
    int type = 0, err = 0;
985 986

    switch (hw_breakpoint_type(env->dr[7], index)) {
987
    case DR7_TYPE_BP_INST:
988
        if (hw_breakpoint_enabled(env->dr[7], index)) {
989 990
            err = cpu_breakpoint_insert(env, env->dr[index], BP_CPU,
                                        &env->cpu_breakpoint[index]);
991
        }
992
        break;
993
    case DR7_TYPE_DATA_WR:
994
        type = BP_CPU | BP_MEM_WRITE;
995
        break;
996
    case DR7_TYPE_IO_RW:
997
        /* No support for I/O watchpoints yet */
998
        break;
999
    case DR7_TYPE_DATA_RW:
1000
        type = BP_CPU | BP_MEM_ACCESS;
1001 1002 1003 1004
        break;
    }

    if (type != 0) {
1005 1006 1007 1008
        err = cpu_watchpoint_insert(env, env->dr[index],
                                    hw_breakpoint_len(env->dr[7], index),
                                    type, &env->cpu_watchpoint[index]);
    }
1009 1010

    if (err) {
1011
        env->cpu_breakpoint[index] = NULL;
1012
    }
1013 1014
}

1015
void hw_breakpoint_remove(CPUX86State *env, int index)
1016 1017 1018 1019
{
    if (!env->cpu_breakpoint[index])
        return;
    switch (hw_breakpoint_type(env->dr[7], index)) {
1020
    case DR7_TYPE_BP_INST:
1021
        if (hw_breakpoint_enabled(env->dr[7], index)) {
1022
            cpu_breakpoint_remove_by_ref(env, env->cpu_breakpoint[index]);
1023
        }
1024
        break;
1025 1026
    case DR7_TYPE_DATA_WR:
    case DR7_TYPE_DATA_RW:
1027 1028
        cpu_watchpoint_remove_by_ref(env, env->cpu_watchpoint[index]);
        break;
1029
    case DR7_TYPE_IO_RW:
1030 1031 1032 1033 1034
        /* No support for I/O watchpoints yet */
        break;
    }
}

1035
bool check_hw_breakpoints(CPUX86State *env, bool force_dr6_update)
1036 1037
{
    target_ulong dr6;
1038 1039
    int reg;
    bool hit_enabled = false;
1040 1041

    dr6 = env->dr[6] & ~0xf;
1042
    for (reg = 0; reg < DR7_MAX_BP; reg++) {
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
        bool bp_match = false;
        bool wp_match = false;

        switch (hw_breakpoint_type(env->dr[7], reg)) {
        case DR7_TYPE_BP_INST:
            if (env->dr[reg] == env->eip) {
                bp_match = true;
            }
            break;
        case DR7_TYPE_DATA_WR:
        case DR7_TYPE_DATA_RW:
            if (env->cpu_watchpoint[reg] &&
                env->cpu_watchpoint[reg]->flags & BP_WATCHPOINT_HIT) {
                wp_match = true;
            }
            break;
        case DR7_TYPE_IO_RW:
            break;
        }
        if (bp_match || wp_match) {
1063
            dr6 |= 1 << reg;
1064
            if (hw_breakpoint_enabled(env->dr[7], reg)) {
1065
                hit_enabled = true;
1066
            }
1067 1068
        }
    }
1069 1070

    if (hit_enabled || force_dr6_update) {
1071
        env->dr[6] = dr6;
1072 1073
    }

1074 1075 1076
    return hit_enabled;
}

1077
void breakpoint_handler(CPUX86State *env)
1078 1079 1080 1081 1082 1083
{
    CPUBreakpoint *bp;

    if (env->watchpoint_hit) {
        if (env->watchpoint_hit->flags & BP_CPU) {
            env->watchpoint_hit = NULL;
1084
            if (check_hw_breakpoints(env, false)) {
B
Blue Swirl 已提交
1085
                raise_exception(env, EXCP01_DB);
1086
            } else {
1087
                cpu_resume_from_signal(env, NULL);
1088
            }
1089 1090
        }
    } else {
B
Blue Swirl 已提交
1091
        QTAILQ_FOREACH(bp, &env->breakpoints, entry)
1092 1093
            if (bp->pc == env->eip) {
                if (bp->flags & BP_CPU) {
1094
                    check_hw_breakpoints(env, true);
B
Blue Swirl 已提交
1095
                    raise_exception(env, EXCP01_DB);
1096 1097 1098 1099 1100
                }
                break;
            }
    }
}
1101

1102 1103
typedef struct MCEInjectionParams {
    Monitor *mon;
1104
    X86CPU *cpu;
1105 1106 1107 1108 1109 1110 1111 1112 1113
    int bank;
    uint64_t status;
    uint64_t mcg_status;
    uint64_t addr;
    uint64_t misc;
    int flags;
} MCEInjectionParams;

static void do_inject_x86_mce(void *data)
1114
{
1115
    MCEInjectionParams *params = data;
1116 1117
    CPUX86State *cenv = &params->cpu->env;
    CPUState *cpu = CPU(params->cpu);
1118 1119
    uint64_t *banks = cenv->mce_banks + 4 * params->bank;

1120
    cpu_synchronize_state(cpu);
1121

1122 1123 1124 1125
    /*
     * If there is an MCE exception being processed, ignore this SRAO MCE
     * unless unconditional injection was requested.
     */
1126 1127
    if (!(params->flags & MCE_INJECT_UNCOND_AO)
        && !(params->status & MCI_STATUS_AR)
1128 1129 1130
        && (cenv->mcg_status & MCG_STATUS_MCIP)) {
        return;
    }
1131 1132

    if (params->status & MCI_STATUS_UC) {
1133 1134 1135 1136
        /*
         * if MSR_MCG_CTL is not all 1s, the uncorrected error
         * reporting is disabled
         */
1137 1138
        if ((cenv->mcg_cap & MCG_CTL_P) && cenv->mcg_ctl != ~(uint64_t)0) {
            monitor_printf(params->mon,
1139
                           "CPU %d: Uncorrected error reporting disabled\n",
1140
                           cpu->cpu_index);
1141 1142 1143 1144 1145 1146 1147 1148
            return;
        }

        /*
         * if MSR_MCi_CTL is not all 1s, the uncorrected error
         * reporting is disabled for the bank
         */
        if (banks[0] != ~(uint64_t)0) {
1149 1150 1151
            monitor_printf(params->mon,
                           "CPU %d: Uncorrected error reporting disabled for"
                           " bank %d\n",
1152
                           cpu->cpu_index, params->bank);
1153 1154 1155
            return;
        }

1156 1157
        if ((cenv->mcg_status & MCG_STATUS_MCIP) ||
            !(cenv->cr[4] & CR4_MCE_MASK)) {
1158 1159 1160
            monitor_printf(params->mon,
                           "CPU %d: Previous MCE still in progress, raising"
                           " triple fault\n",
1161
                           cpu->cpu_index);
1162 1163 1164 1165
            qemu_log_mask(CPU_LOG_RESET, "Triple fault\n");
            qemu_system_reset_request();
            return;
        }
J
Jan Kiszka 已提交
1166
        if (banks[1] & MCI_STATUS_VAL) {
1167
            params->status |= MCI_STATUS_OVER;
J
Jan Kiszka 已提交
1168
        }
1169 1170 1171 1172
        banks[2] = params->addr;
        banks[3] = params->misc;
        cenv->mcg_status = params->mcg_status;
        banks[1] = params->status;
1173
        cpu_interrupt(cpu, CPU_INTERRUPT_MCE);
1174 1175
    } else if (!(banks[1] & MCI_STATUS_VAL)
               || !(banks[1] & MCI_STATUS_UC)) {
J
Jan Kiszka 已提交
1176
        if (banks[1] & MCI_STATUS_VAL) {
1177
            params->status |= MCI_STATUS_OVER;
J
Jan Kiszka 已提交
1178
        }
1179 1180 1181
        banks[2] = params->addr;
        banks[3] = params->misc;
        banks[1] = params->status;
J
Jan Kiszka 已提交
1182
    } else {
1183
        banks[1] |= MCI_STATUS_OVER;
J
Jan Kiszka 已提交
1184
    }
1185
}
J
Jin Dongming 已提交
1186

1187
void cpu_x86_inject_mce(Monitor *mon, X86CPU *cpu, int bank,
1188
                        uint64_t status, uint64_t mcg_status, uint64_t addr,
1189
                        uint64_t misc, int flags)
J
Jin Dongming 已提交
1190
{
1191
    CPUState *cs = CPU(cpu);
1192
    CPUX86State *cenv = &cpu->env;
1193 1194
    MCEInjectionParams params = {
        .mon = mon,
1195
        .cpu = cpu,
1196 1197 1198 1199 1200 1201 1202
        .bank = bank,
        .status = status,
        .mcg_status = mcg_status,
        .addr = addr,
        .misc = misc,
        .flags = flags,
    };
J
Jin Dongming 已提交
1203 1204
    unsigned bank_num = cenv->mcg_cap & 0xff;

1205 1206
    if (!cenv->mcg_cap) {
        monitor_printf(mon, "MCE injection not supported\n");
J
Jin Dongming 已提交
1207 1208
        return;
    }
1209 1210 1211 1212 1213 1214 1215 1216
    if (bank >= bank_num) {
        monitor_printf(mon, "Invalid MCE bank number\n");
        return;
    }
    if (!(status & MCI_STATUS_VAL)) {
        monitor_printf(mon, "Invalid MCE status code\n");
        return;
    }
1217 1218
    if ((flags & MCE_INJECT_BROADCAST)
        && !cpu_x86_support_mca_broadcast(cenv)) {
1219 1220
        monitor_printf(mon, "Guest CPU does not support MCA broadcast\n");
        return;
1221 1222
    }

1223
    run_on_cpu(cs, do_inject_x86_mce, &params);
1224
    if (flags & MCE_INJECT_BROADCAST) {
1225 1226
        CPUState *other_cs;

1227 1228 1229 1230 1231
        params.bank = 1;
        params.status = MCI_STATUS_VAL | MCI_STATUS_UC;
        params.mcg_status = MCG_STATUS_MCIP | MCG_STATUS_RIPV;
        params.addr = 0;
        params.misc = 0;
A
Andreas Färber 已提交
1232
        CPU_FOREACH(other_cs) {
1233
            if (other_cs == cs) {
1234
                continue;
1235
            }
1236 1237
            params.cpu = X86_CPU(other_cs);
            run_on_cpu(other_cs, do_inject_x86_mce, &params);
1238
        }
J
Jin Dongming 已提交
1239 1240
    }
}
1241

1242
void cpu_report_tpr_access(CPUX86State *env, TPRAccess access)
1243 1244 1245 1246
{
    if (kvm_enabled()) {
        env->tpr_access_type = access;

1247
        cpu_interrupt(CPU(x86_env_get_cpu(env)), CPU_INTERRUPT_TPR);
1248
    } else {
B
Blue Swirl 已提交
1249
        cpu_restore_state(env, env->mem_io_pc);
1250 1251 1252 1253

        apic_handle_tpr_access_report(env->apic_state, env->eip, access);
    }
}
B
bellard 已提交
1254
#endif /* !CONFIG_USER_ONLY */
A
aliguori 已提交
1255

1256 1257 1258 1259
int cpu_x86_get_descr_debug(CPUX86State *env, unsigned int selector,
                            target_ulong *base, unsigned int *limit,
                            unsigned int *flags)
{
1260 1261
    X86CPU *cpu = x86_env_get_cpu(env);
    CPUState *cs = CPU(cpu);
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
    SegmentCache *dt;
    target_ulong ptr;
    uint32_t e1, e2;
    int index;

    if (selector & 0x4)
        dt = &env->ldt;
    else
        dt = &env->gdt;
    index = selector & ~7;
    ptr = dt->base + index;
    if ((index + 7) > dt->limit
1274 1275
        || cpu_memory_rw_debug(cs, ptr, (uint8_t *)&e1, sizeof(e1), 0) != 0
        || cpu_memory_rw_debug(cs, ptr+4, (uint8_t *)&e2, sizeof(e2), 0) != 0)
1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286
        return 0;

    *base = ((e1 >> 16) | ((e2 & 0xff) << 16) | (e2 & 0xff000000));
    *limit = (e1 & 0xffff) | (e2 & 0x000f0000);
    if (e2 & DESC_G_MASK)
        *limit = (*limit << 12) | 0xfff;
    *flags = e2;

    return 1;
}

1287
#if !defined(CONFIG_USER_ONLY)
1288
void do_cpu_init(X86CPU *cpu)
1289
{
1290
    CPUState *cs = CPU(cpu);
1291
    CPUX86State *env = &cpu->env;
1292
    int sipi = cs->interrupt_request & CPU_INTERRUPT_SIPI;
J
Jan Kiszka 已提交
1293 1294
    uint64_t pat = env->pat;

1295 1296
    cpu_reset(cs);
    cs->interrupt_request = sipi;
J
Jan Kiszka 已提交
1297
    env->pat = pat;
1298
    apic_init_reset(env->apic_state);
1299 1300
}

1301
void do_cpu_sipi(X86CPU *cpu)
1302
{
1303 1304
    CPUX86State *env = &cpu->env;

1305
    apic_sipi(env->apic_state);
1306 1307
}
#else
1308
void do_cpu_init(X86CPU *cpu)
1309 1310
{
}
1311
void do_cpu_sipi(X86CPU *cpu)
1312 1313 1314
{
}
#endif