vl.c 157.5 KB
Newer Older
1
/*
B
bellard 已提交
2
 * QEMU System Emulator
3
 *
B
bellard 已提交
4
 * Copyright (c) 2003-2008 Fabrice Bellard
5
 *
B
bellard 已提交
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
 * 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.
23 24 25 26 27 28
 */
#include <unistd.h>
#include <fcntl.h>
#include <signal.h>
#include <time.h>
#include <errno.h>
B
bellard 已提交
29
#include <sys/time.h>
30
#include <zlib.h>
B
bellard 已提交
31

J
Juan Quintela 已提交
32
/* Needed early for CONFIG_BSD etc. */
B
blueswir1 已提交
33 34
#include "config-host.h"

B
bellard 已提交
35
#ifndef _WIN32
P
Paul Brook 已提交
36
#include <libgen.h>
37
#include <pwd.h>
B
bellard 已提交
38
#include <sys/times.h>
B
bellard 已提交
39
#include <sys/wait.h>
B
bellard 已提交
40 41
#include <termios.h>
#include <sys/mman.h>
B
bellard 已提交
42
#include <sys/ioctl.h>
B
blueswir1 已提交
43
#include <sys/resource.h>
B
bellard 已提交
44
#include <sys/socket.h>
B
bellard 已提交
45
#include <netinet/in.h>
B
blueswir1 已提交
46 47
#include <net/if.h>
#include <arpa/inet.h>
B
bellard 已提交
48
#include <dirent.h>
B
bellard 已提交
49
#include <netdb.h>
50
#include <sys/select.h>
J
Juan Quintela 已提交
51
#ifdef CONFIG_BSD
B
bellard 已提交
52
#include <sys/stat.h>
A
Aurelien Jarno 已提交
53
#if defined(__FreeBSD__) || defined(__FreeBSD_kernel__) || defined(__DragonFly__)
B
bellard 已提交
54
#include <libutil.h>
B
blueswir1 已提交
55 56
#else
#include <util.h>
57
#endif
58
#else
59
#ifdef __linux__
B
bellard 已提交
60 61
#include <pty.h>
#include <malloc.h>
62
#include <linux/rtc.h>
A
Andi Kleen 已提交
63
#include <sys/prctl.h>
T
ths 已提交
64 65 66 67 68 69

/* For the benefit of older linux systems which don't supply it,
   we use a local copy of hpet.h. */
/* #include <linux/hpet.h> */
#include "hpet.h"

B
bellard 已提交
70
#include <linux/ppdev.h>
71
#include <linux/parport.h>
72 73
#endif
#ifdef __sun__
74 75 76 77 78 79 80 81 82 83 84 85 86
#include <sys/stat.h>
#include <sys/ethernet.h>
#include <sys/sockio.h>
#include <netinet/arp.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/ip_icmp.h> // must come after ip.h
#include <netinet/udp.h>
#include <netinet/tcp.h>
#include <net/if.h>
#include <syslog.h>
#include <stropts.h>
87 88 89
/* See MySQL bug #7156 (http://bugs.mysql.com/bug.php?id=7156) for
   discussion about Solaris header problems */
extern int madvise(caddr_t, size_t, int);
B
bellard 已提交
90
#endif
B
bellard 已提交
91
#endif
B
bellard 已提交
92
#endif
B
bellard 已提交
93

94 95 96 97
#if defined(__OpenBSD__)
#include <util.h>
#endif

98 99 100 101
#if defined(CONFIG_VDE)
#include <libvdeplug.h>
#endif

B
bellard 已提交
102
#ifdef _WIN32
103
#include <windows.h>
T
ths 已提交
104
#include <mmsystem.h>
B
bellard 已提交
105 106
#endif

B
bellard 已提交
107
#ifdef CONFIG_SDL
S
Stefan Weil 已提交
108
#if defined(__APPLE__) || defined(main)
S
Stefan Weil 已提交
109
#include <SDL.h>
M
malc 已提交
110 111 112
int qemu_main(int argc, char **argv, char **envp);
int main(int argc, char **argv)
{
S
Stefan Weil 已提交
113
    return qemu_main(argc, argv, NULL);
M
malc 已提交
114 115 116
}
#undef main
#define main qemu_main
B
bellard 已提交
117
#endif
B
bellard 已提交
118
#endif /* CONFIG_SDL */
119

120 121 122 123 124
#ifdef CONFIG_COCOA
#undef main
#define main qemu_main
#endif /* CONFIG_COCOA */

125 126 127 128 129 130 131 132 133
#include "hw/hw.h"
#include "hw/boards.h"
#include "hw/usb.h"
#include "hw/pcmcia.h"
#include "hw/pc.h"
#include "hw/audiodev.h"
#include "hw/isa.h"
#include "hw/baum.h"
#include "hw/bt.h"
R
Richard W.M. Jones 已提交
134
#include "hw/watchdog.h"
135
#include "hw/smbios.h"
136
#include "hw/xen.h"
137
#include "hw/qdev.h"
138
#include "hw/loader.h"
A
aurel32 已提交
139
#include "bt-host.h"
140
#include "net.h"
141
#include "net/slirp.h"
142 143 144 145 146 147 148 149
#include "monitor.h"
#include "console.h"
#include "sysemu.h"
#include "gdbstub.h"
#include "qemu-timer.h"
#include "qemu-char.h"
#include "cache-utils.h"
#include "block.h"
L
lirans@il.ibm.com 已提交
150 151
#include "block_int.h"
#include "block-migration.h"
B
blueswir1 已提交
152
#include "dma.h"
153 154 155 156
#include "audio/audio.h"
#include "migration.h"
#include "kvm.h"
#include "balloon.h"
K
Kevin Wolf 已提交
157
#include "qemu-option.h"
G
Gerd Hoffmann 已提交
158
#include "qemu-config.h"
159
#include "qemu-objects.h"
160

161
#include "disas.h"
B
bellard 已提交
162

163
#include "exec-all.h"
164

165 166
#include "qemu_socket.h"

J
Jan Kiszka 已提交
167
#include "slirp/libslirp.h"
168

B
Blue Swirl 已提交
169 170
#include "qemu-queue.h"

171 172
//#define DEBUG_NET
//#define DEBUG_SLIRP
173

174
#define DEFAULT_RAM_SIZE 128
175

176 177
#define MAX_VIRTIO_CONSOLES 1

P
Paul Brook 已提交
178
static const char *data_dir;
179
const char *bios_name = NULL;
T
ths 已提交
180
/* Note: drives_table[MAX_DRIVES] is a dummy block driver if none available
B
bellard 已提交
181
   to store the VM snapshots */
B
Blue Swirl 已提交
182 183
struct drivelist drives = QTAILQ_HEAD_INITIALIZER(drives);
struct driveoptlist driveopts = QTAILQ_HEAD_INITIALIZER(driveopts);
M
malc 已提交
184
enum vga_retrace_method vga_retrace_method = VGA_RETRACE_DUMB;
185
DisplayType display_type = DT_DEFAULT;
186
const char* keyboard_layout = NULL;
A
Anthony Liguori 已提交
187
ram_addr_t ram_size;
188 189 190 191
const char *mem_path = NULL;
#ifdef MAP_POPULATE
int mem_prealloc = 0; /* force preallocation of physical target memory */
#endif
192
int nb_nics;
B
bellard 已提交
193
NICInfo nd_table[MAX_NICS];
194
int vm_running;
P
Paolo Bonzini 已提交
195
int autostart;
196 197
static int rtc_utc = 1;
static int rtc_date_offset = -1; /* -1 means no change */
J
Jan Kiszka 已提交
198
QEMUClock *rtc_clock;
G
Gerd Hoffmann 已提交
199
int vga_interface_type = VGA_NONE;
200 201 202
#ifdef TARGET_SPARC
int graphic_width = 1024;
int graphic_height = 768;
B
blueswir1 已提交
203
int graphic_depth = 8;
204
#else
205 206
int graphic_width = 800;
int graphic_height = 600;
207
int graphic_depth = 15;
B
blueswir1 已提交
208
#endif
B
blueswir1 已提交
209
static int full_screen = 0;
210
#ifdef CONFIG_SDL
B
blueswir1 已提交
211
static int no_frame = 0;
212
#endif
T
ths 已提交
213
int no_quit = 0;
214
CharDriverState *serial_hds[MAX_SERIAL_PORTS];
215
CharDriverState *parallel_hds[MAX_PARALLEL_PORTS];
216
CharDriverState *virtcon_hds[MAX_VIRTIO_CONSOLES];
217 218
#ifdef TARGET_I386
int win2k_install_hack = 0;
219
int rtc_td_hack = 0;
220
#endif
B
bellard 已提交
221
int usb_enabled = 0;
222
int singlestep = 0;
B
bellard 已提交
223
int smp_cpus = 1;
224
int max_cpus = 0;
225 226
int smp_cores = 1;
int smp_threads = 1;
227
const char *vnc_display;
B
bellard 已提交
228
int acpi_enabled = 1;
A
aliguori 已提交
229
int no_hpet = 0;
B
bellard 已提交
230
int fd_bootchk = 1;
B
bellard 已提交
231
int no_reboot = 0;
A
aurel32 已提交
232
int no_shutdown = 0;
233
int cursor_hide = 1;
234
int graphic_rotate = 0;
J
Jes Sorensen 已提交
235
uint8_t irq0override = 1;
B
blueswir1 已提交
236
#ifndef _WIN32
T
ths 已提交
237
int daemonize = 0;
B
blueswir1 已提交
238
#endif
M
Markus Armbruster 已提交
239
const char *watchdog;
240 241
const char *option_rom[MAX_OPTION_ROMS];
int nb_option_roms;
242
int semihosting_enabled = 0;
243 244 245
#ifdef TARGET_ARM
int old_param = 0;
#endif
T
ths 已提交
246
const char *qemu_name;
T
ths 已提交
247
int alt_grab = 0;
248
int ctrl_grab = 0;
249
#if defined(TARGET_SPARC) || defined(TARGET_PPC)
B
blueswir1 已提交
250 251 252
unsigned int nb_prom_envs = 0;
const char *prom_envs[MAX_PROM_ENVS];
#endif
253
int boot_menu;
254

255 256 257 258
int nb_numa_nodes;
uint64_t node_mem[MAX_NODES];
uint64_t node_cpumask[MAX_NODES];

259 260
static CPUState *cur_cpu;
static CPUState *next_cpu;
T
ths 已提交
261
/* Conversion factor from emulated instructions to virtual clock ticks.  */
P
pbrook 已提交
262
static int icount_time_shift;
T
ths 已提交
263
/* Arbitrarily pick 1MIPS as the minimum allowable speed.  */
P
pbrook 已提交
264 265 266
#define MAX_ICOUNT_SHIFT 10
/* Compensate for varying guest execution speed.  */
static int64_t qemu_icount_bias;
B
blueswir1 已提交
267 268
static QEMUTimer *icount_rt_timer;
static QEMUTimer *icount_vm_timer;
B
blueswir1 已提交
269
static QEMUTimer *nographic_timer;
270

271 272
uint8_t qemu_uuid[16];

J
Jan Kiszka 已提交
273 274 275
static QEMUBootSetHandler *boot_set_handler;
static void *boot_set_opaque;

276 277 278 279 280 281
#ifdef SIGRTMIN
#define SIG_IPI (SIGRTMIN+4)
#else
#define SIG_IPI SIGUSR1
#endif

282
static int default_serial = 1;
G
Gerd Hoffmann 已提交
283
static int default_parallel = 1;
284
static int default_virtcon = 1;
G
Gerd Hoffmann 已提交
285
static int default_monitor = 1;
G
Gerd Hoffmann 已提交
286
static int default_vga = 1;
G
Gerd Hoffmann 已提交
287 288 289
static int default_floppy = 1;
static int default_cdrom = 1;
static int default_sdcard = 1;
290 291 292 293 294

static struct {
    const char *driver;
    int *flag;
} default_list[] = {
G
Gerd Hoffmann 已提交
295 296
    { .driver = "isa-serial",           .flag = &default_serial    },
    { .driver = "isa-parallel",         .flag = &default_parallel  },
G
Gerd Hoffmann 已提交
297 298
    { .driver = "isa-fdc",              .flag = &default_floppy    },
    { .driver = "ide-drive",            .flag = &default_cdrom     },
299 300 301
    { .driver = "virtio-serial-pci",    .flag = &default_virtcon   },
    { .driver = "virtio-serial-s390",   .flag = &default_virtcon   },
    { .driver = "virtio-serial",        .flag = &default_virtcon   },
G
Gerd Hoffmann 已提交
302
    { .driver = "VGA",                  .flag = &default_vga       },
G
Gerd Hoffmann 已提交
303 304
    { .driver = "cirrus-vga",           .flag = &default_vga       },
    { .driver = "vmware-svga",          .flag = &default_vga       },
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
};

static int default_driver_check(QemuOpts *opts, void *opaque)
{
    const char *driver = qemu_opt_get(opts, "driver");
    int i;

    if (!driver)
        return 0;
    for (i = 0; i < ARRAY_SIZE(default_list); i++) {
        if (strcmp(default_list[i].driver, driver) != 0)
            continue;
        *(default_list[i].flag) = 0;
    }
    return 0;
}

322
/***********************************************************/
323 324
/* x86 ISA bus support */

A
Anthony Liguori 已提交
325
target_phys_addr_t isa_mem_base = 0;
B
bellard 已提交
326
PicState2 *isa_pic;
327 328 329 330 331

/***********************************************************/
void hw_error(const char *fmt, ...)
{
    va_list ap;
B
bellard 已提交
332
    CPUState *env;
333 334 335 336 337

    va_start(ap, fmt);
    fprintf(stderr, "qemu: hardware error: ");
    vfprintf(stderr, fmt, ap);
    fprintf(stderr, "\n");
B
bellard 已提交
338 339
    for(env = first_cpu; env != NULL; env = env->next_cpu) {
        fprintf(stderr, "CPU #%d:\n", env->cpu_index);
340
#ifdef TARGET_I386
B
bellard 已提交
341
        cpu_dump_state(env, stderr, fprintf, X86_DUMP_FPU);
342
#else
B
bellard 已提交
343
        cpu_dump_state(env, stderr, fprintf, 0);
344
#endif
B
bellard 已提交
345
    }
346 347 348
    va_end(ap);
    abort();
}
A
Andi Kleen 已提交
349 350 351

static void set_proc_name(const char *s)
{
352
#if defined(__linux__) && defined(PR_SET_NAME)
A
Andi Kleen 已提交
353 354 355 356 357 358 359 360 361 362
    char name[16];
    if (!s)
        return;
    name[sizeof(name) - 1] = 0;
    strncpy(name, s, sizeof(name));
    /* Could rewrite argv[0] too, but that's a bit more complicated.
       This simple way is enough for `top'. */
    prctl(PR_SET_NAME, name);
#endif    	
}
A
aliguori 已提交
363 364 365 366 367 368 369 370 371 372 373 374 375
 
/***************/
/* ballooning */

static QEMUBalloonEvent *qemu_balloon_event;
void *qemu_balloon_event_opaque;

void qemu_add_balloon_handler(QEMUBalloonEvent *func, void *opaque)
{
    qemu_balloon_event = func;
    qemu_balloon_event_opaque = opaque;
}

376
int qemu_balloon(ram_addr_t target, MonitorCompletion cb, void *opaque)
A
aliguori 已提交
377
{
378 379 380 381 382 383
    if (qemu_balloon_event) {
        qemu_balloon_event(qemu_balloon_event_opaque, target, cb, opaque);
        return 1;
    } else {
        return 0;
    }
A
aliguori 已提交
384 385
}

386
int qemu_balloon_status(MonitorCompletion cb, void *opaque)
A
aliguori 已提交
387
{
388 389 390 391 392 393
    if (qemu_balloon_event) {
        qemu_balloon_event(qemu_balloon_event_opaque, 0, cb, opaque);
        return 1;
    } else {
        return 0;
    }
A
aliguori 已提交
394
}
395

396

397 398
/***********************************************************/
/* real time host monotonic timer */
399

400 401
/* compute with 96 bit intermediate result: (a*b)/c */
uint64_t muldiv64(uint64_t a, uint32_t b, uint32_t c)
402
{
403 404 405
    union {
        uint64_t ll;
        struct {
406
#ifdef HOST_WORDS_BIGENDIAN
407 408 409
            uint32_t high, low;
#else
            uint32_t low, high;
410
#endif
411 412 413
        } l;
    } u, res;
    uint64_t rl, rh;
414

415 416 417 418 419 420 421
    u.ll = a;
    rl = (uint64_t)u.l.low * (uint64_t)b;
    rh = (uint64_t)u.l.high * (uint64_t)b;
    rh += (rl >> 32);
    res.l.high = rh / c;
    res.l.low = (((rh % c) << 32) + (rl & 0xffffffff)) / c;
    return res.ll;
422 423
}

J
Jan Kiszka 已提交
424 425 426 427 428 429 430 431
static int64_t get_clock_realtime(void)
{
    struct timeval tv;

    gettimeofday(&tv, NULL);
    return tv.tv_sec * 1000000000LL + (tv.tv_usec * 1000);
}

432
#ifdef WIN32
433

434
static int64_t clock_freq;
B
bellard 已提交
435

436
static void init_get_clock(void)
B
bellard 已提交
437
{
438 439
    LARGE_INTEGER freq;
    int ret;
440 441 442 443 444 445
    ret = QueryPerformanceFrequency(&freq);
    if (ret == 0) {
        fprintf(stderr, "Could not calibrate ticks\n");
        exit(1);
    }
    clock_freq = freq.QuadPart;
B
bellard 已提交
446 447
}

448
static int64_t get_clock(void)
B
bellard 已提交
449
{
450 451
    LARGE_INTEGER ti;
    QueryPerformanceCounter(&ti);
452
    return muldiv64(ti.QuadPart, get_ticks_per_sec(), clock_freq);
B
bellard 已提交
453 454
}

455
#else
B
bellard 已提交
456

457 458 459
static int use_rt_clock;

static void init_get_clock(void)
B
bellard 已提交
460
{
461
    use_rt_clock = 0;
462
#if defined(__linux__) || (defined(__FreeBSD__) && __FreeBSD_version >= 500000) \
A
Aurelien Jarno 已提交
463
    || defined(__DragonFly__) || defined(__FreeBSD_kernel__)
464 465 466 467 468 469 470
    {
        struct timespec ts;
        if (clock_gettime(CLOCK_MONOTONIC, &ts) == 0) {
            use_rt_clock = 1;
        }
    }
#endif
B
bellard 已提交
471 472
}

473
static int64_t get_clock(void)
B
bellard 已提交
474
{
475
#if defined(__linux__) || (defined(__FreeBSD__) && __FreeBSD_version >= 500000) \
A
Aurelien Jarno 已提交
476
	|| defined(__DragonFly__) || defined(__FreeBSD_kernel__)
477 478 479 480
    if (use_rt_clock) {
        struct timespec ts;
        clock_gettime(CLOCK_MONOTONIC, &ts);
        return ts.tv_sec * 1000000000LL + ts.tv_nsec;
481
    } else
B
bellard 已提交
482
#endif
483 484 485
    {
        /* XXX: using gettimeofday leads to problems if the date
           changes, so it should be avoided. */
J
Jan Kiszka 已提交
486
        return get_clock_realtime();
487
    }
B
bellard 已提交
488
}
489 490
#endif

P
pbrook 已提交
491 492 493 494 495 496 497 498 499 500 501 502 503 504
/* Return the virtual CPU time, based on the instruction counter.  */
static int64_t cpu_get_icount(void)
{
    int64_t icount;
    CPUState *env = cpu_single_env;;
    icount = qemu_icount;
    if (env) {
        if (!can_do_io(env))
            fprintf(stderr, "Bad clock read\n");
        icount -= (env->icount_decr.u16.low + env->icount_extra);
    }
    return qemu_icount_bias + (icount << icount_time_shift);
}

505 506 507
/***********************************************************/
/* guest cycle counter */

508 509 510 511 512
typedef struct TimersState {
    int64_t cpu_ticks_prev;
    int64_t cpu_ticks_offset;
    int64_t cpu_clock_offset;
    int32_t cpu_ticks_enabled;
513
    int64_t dummy;
514 515 516
} TimersState;

TimersState timers_state;
517

518 519
/* return the host CPU cycle counter and handle stop/restart */
int64_t cpu_get_ticks(void)
520
{
P
pbrook 已提交
521 522 523
    if (use_icount) {
        return cpu_get_icount();
    }
524 525
    if (!timers_state.cpu_ticks_enabled) {
        return timers_state.cpu_ticks_offset;
526
    } else {
527 528
        int64_t ticks;
        ticks = cpu_get_real_ticks();
529
        if (timers_state.cpu_ticks_prev > ticks) {
530 531
            /* Note: non increasing ticks may happen if the host uses
               software suspend */
532
            timers_state.cpu_ticks_offset += timers_state.cpu_ticks_prev - ticks;
533
        }
534 535
        timers_state.cpu_ticks_prev = ticks;
        return ticks + timers_state.cpu_ticks_offset;
536
    }
537 538
}

539 540 541 542
/* return the host CPU monotonic timer and handle stop/restart */
static int64_t cpu_get_clock(void)
{
    int64_t ti;
543 544
    if (!timers_state.cpu_ticks_enabled) {
        return timers_state.cpu_clock_offset;
545 546
    } else {
        ti = get_clock();
547
        return ti + timers_state.cpu_clock_offset;
548 549 550
    }
}

551 552 553
/* enable cpu_get_ticks() */
void cpu_enable_ticks(void)
{
554 555 556 557
    if (!timers_state.cpu_ticks_enabled) {
        timers_state.cpu_ticks_offset -= cpu_get_real_ticks();
        timers_state.cpu_clock_offset -= get_clock();
        timers_state.cpu_ticks_enabled = 1;
558
    }
559 560 561 562 563 564
}

/* disable cpu_get_ticks() : the clock is stopped. You must not call
   cpu_get_ticks() after that.  */
void cpu_disable_ticks(void)
{
565 566 567 568
    if (timers_state.cpu_ticks_enabled) {
        timers_state.cpu_ticks_offset = cpu_get_ticks();
        timers_state.cpu_clock_offset = cpu_get_clock();
        timers_state.cpu_ticks_enabled = 0;
569
    }
570 571
}

572 573
/***********************************************************/
/* timers */
574

J
Jan Kiszka 已提交
575 576
#define QEMU_CLOCK_REALTIME 0
#define QEMU_CLOCK_VIRTUAL  1
J
Jan Kiszka 已提交
577
#define QEMU_CLOCK_HOST     2
578 579 580

struct QEMUClock {
    int type;
581
    int enabled;
582 583 584 585 586 587 588 589 590 591 592
    /* XXX: add frequency */
};

struct QEMUTimer {
    QEMUClock *clock;
    int64_t expire_time;
    QEMUTimerCB *cb;
    void *opaque;
    struct QEMUTimer *next;
};

593 594 595 596
struct qemu_alarm_timer {
    char const *name;
    int (*start)(struct qemu_alarm_timer *t);
    void (*stop)(struct qemu_alarm_timer *t);
597
    void (*rearm)(struct qemu_alarm_timer *t);
598 599
    void *priv;

P
Paolo Bonzini 已提交
600 601
    char expired;
    char pending;
602
};
603

P
Paolo Bonzini 已提交
604 605 606 607 608 609 610
static struct qemu_alarm_timer *alarm_timer;

static inline int qemu_alarm_pending(void)
{
    return alarm_timer->pending;
}

611 612
static inline int alarm_has_dynticks(struct qemu_alarm_timer *t)
{
P
Paolo Bonzini 已提交
613
    return !!t->rearm;
614 615 616 617 618 619 620 621 622 623 624 625 626
}

static void qemu_rearm_alarm_timer(struct qemu_alarm_timer *t)
{
    if (!alarm_has_dynticks(t))
        return;

    t->rearm(t);
}

/* TODO: MIN_TIMER_REARM_US should be optimized */
#define MIN_TIMER_REARM_US 250

B
bellard 已提交
627
#ifdef _WIN32
628 629 630 631

struct qemu_alarm_win32 {
    MMRESULT timerId;
    unsigned int period;
632
} alarm_win32_data = {0, 0};
633 634 635

static int win32_start_timer(struct qemu_alarm_timer *t);
static void win32_stop_timer(struct qemu_alarm_timer *t);
636
static void win32_rearm_timer(struct qemu_alarm_timer *t);
637

B
bellard 已提交
638
#else
639 640 641 642

static int unix_start_timer(struct qemu_alarm_timer *t);
static void unix_stop_timer(struct qemu_alarm_timer *t);

643 644
#ifdef __linux__

645 646 647 648
static int dynticks_start_timer(struct qemu_alarm_timer *t);
static void dynticks_stop_timer(struct qemu_alarm_timer *t);
static void dynticks_rearm_timer(struct qemu_alarm_timer *t);

T
ths 已提交
649 650 651
static int hpet_start_timer(struct qemu_alarm_timer *t);
static void hpet_stop_timer(struct qemu_alarm_timer *t);

652 653 654
static int rtc_start_timer(struct qemu_alarm_timer *t);
static void rtc_stop_timer(struct qemu_alarm_timer *t);

655
#endif /* __linux__ */
656

657 658
#endif /* _WIN32 */

P
pbrook 已提交
659
/* Correlation between real and virtual time is always going to be
T
ths 已提交
660
   fairly approximate, so ignore small variation.
P
pbrook 已提交
661 662
   When the guest is idle real and virtual time will be aligned in
   the IO wait loop.  */
663
#define ICOUNT_WOBBLE (get_ticks_per_sec() / 10)
P
pbrook 已提交
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704

static void icount_adjust(void)
{
    int64_t cur_time;
    int64_t cur_icount;
    int64_t delta;
    static int64_t last_delta;
    /* If the VM is not running, then do nothing.  */
    if (!vm_running)
        return;

    cur_time = cpu_get_clock();
    cur_icount = qemu_get_clock(vm_clock);
    delta = cur_icount - cur_time;
    /* FIXME: This is a very crude algorithm, somewhat prone to oscillation.  */
    if (delta > 0
        && last_delta + ICOUNT_WOBBLE < delta * 2
        && icount_time_shift > 0) {
        /* The guest is getting too far ahead.  Slow time down.  */
        icount_time_shift--;
    }
    if (delta < 0
        && last_delta - ICOUNT_WOBBLE > delta * 2
        && icount_time_shift < MAX_ICOUNT_SHIFT) {
        /* The guest is getting too far behind.  Speed time up.  */
        icount_time_shift++;
    }
    last_delta = delta;
    qemu_icount_bias = cur_icount - (qemu_icount << icount_time_shift);
}

static void icount_adjust_rt(void * opaque)
{
    qemu_mod_timer(icount_rt_timer,
                   qemu_get_clock(rt_clock) + 1000);
    icount_adjust();
}

static void icount_adjust_vm(void * opaque)
{
    qemu_mod_timer(icount_vm_timer,
705
                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
P
pbrook 已提交
706 707 708
    icount_adjust();
}

P
Paolo Bonzini 已提交
709
static void configure_icount(const char *option)
P
pbrook 已提交
710
{
P
Paolo Bonzini 已提交
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725
    if (!option)
        return;

    if (strcmp(option, "auto") != 0) {
        icount_time_shift = strtol(option, NULL, 0);
        use_icount = 1;
        return;
    }

    use_icount = 2;

    /* 125MIPS seems a reasonable initial guess at the guest speed.
       It will be corrected fairly quickly anyway.  */
    icount_time_shift = 3;

P
pbrook 已提交
726 727 728 729 730 731 732 733 734 735
    /* Have both realtime and virtual time triggers for speed adjustment.
       The realtime trigger catches emulated time passing too slowly,
       the virtual time trigger catches emulated time passing too fast.
       Realtime triggers occur even when idle, so use them less frequently
       than VM triggers.  */
    icount_rt_timer = qemu_new_timer(rt_clock, icount_adjust_rt, NULL);
    qemu_mod_timer(icount_rt_timer,
                   qemu_get_clock(rt_clock) + 1000);
    icount_vm_timer = qemu_new_timer(vm_clock, icount_adjust_vm, NULL);
    qemu_mod_timer(icount_vm_timer,
736
                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
P
pbrook 已提交
737 738
}

P
Paolo Bonzini 已提交
739 740 741 742 743
static int64_t qemu_icount_round(int64_t count)
{
    return (count + (1 << icount_time_shift) - 1) >> icount_time_shift;
}

744
static struct qemu_alarm_timer alarm_timers[] = {
745
#ifndef _WIN32
746
#ifdef __linux__
747
    {"dynticks", dynticks_start_timer,
748
     dynticks_stop_timer, dynticks_rearm_timer, NULL},
T
ths 已提交
749
    /* HPET - if available - is preferred */
750
    {"hpet", hpet_start_timer, hpet_stop_timer, NULL, NULL},
T
ths 已提交
751
    /* ...otherwise try RTC */
752
    {"rtc", rtc_start_timer, rtc_stop_timer, NULL, NULL},
753
#endif
754
    {"unix", unix_start_timer, unix_stop_timer, NULL, NULL},
755
#else
756
    {"dynticks", win32_start_timer,
757
     win32_stop_timer, win32_rearm_timer, &alarm_win32_data},
758
    {"win32", win32_start_timer,
759
     win32_stop_timer, NULL, &alarm_win32_data},
760 761 762 763
#endif
    {NULL, }
};

764
static void show_available_alarms(void)
765 766 767 768 769 770 771 772 773 774 775 776
{
    int i;

    printf("Available alarm timers, in order of precedence:\n");
    for (i = 0; alarm_timers[i].name; i++)
        printf("%s\n", alarm_timers[i].name);
}

static void configure_alarms(char const *opt)
{
    int i;
    int cur = 0;
777
    int count = ARRAY_SIZE(alarm_timers) - 1;
778 779
    char *arg;
    char *name;
P
pbrook 已提交
780
    struct qemu_alarm_timer tmp;
781

782
    if (!strcmp(opt, "?")) {
783 784 785 786
        show_available_alarms();
        exit(0);
    }

787
    arg = qemu_strdup(opt);
788 789 790 791

    /* Reorder the array */
    name = strtok(arg, ",");
    while (name) {
792
        for (i = 0; i < count && alarm_timers[i].name; i++) {
793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
            if (!strcmp(alarm_timers[i].name, name))
                break;
        }

        if (i == count) {
            fprintf(stderr, "Unknown clock %s\n", name);
            goto next;
        }

        if (i < cur)
            /* Ignore */
            goto next;

	/* Swap */
        tmp = alarm_timers[i];
        alarm_timers[i] = alarm_timers[cur];
        alarm_timers[cur] = tmp;

        cur++;
next:
        name = strtok(NULL, ",");
    }

816
    qemu_free(arg);
817 818

    if (cur) {
P
pbrook 已提交
819
        /* Disable remaining timers */
820 821
        for (i = cur; i < count; i++)
            alarm_timers[i].name = NULL;
822 823 824
    } else {
        show_available_alarms();
        exit(1);
825 826 827
    }
}

J
Jan Kiszka 已提交
828 829
#define QEMU_NUM_CLOCKS 3

830 831
QEMUClock *rt_clock;
QEMUClock *vm_clock;
J
Jan Kiszka 已提交
832
QEMUClock *host_clock;
833

J
Jan Kiszka 已提交
834
static QEMUTimer *active_timers[QEMU_NUM_CLOCKS];
835

836
static QEMUClock *qemu_new_clock(int type)
837 838 839 840
{
    QEMUClock *clock;
    clock = qemu_mallocz(sizeof(QEMUClock));
    clock->type = type;
841
    clock->enabled = 1;
842 843 844
    return clock;
}

845 846 847 848 849
static void qemu_clock_enable(QEMUClock *clock, int enabled)
{
    clock->enabled = enabled;
}

850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901
QEMUTimer *qemu_new_timer(QEMUClock *clock, QEMUTimerCB *cb, void *opaque)
{
    QEMUTimer *ts;

    ts = qemu_mallocz(sizeof(QEMUTimer));
    ts->clock = clock;
    ts->cb = cb;
    ts->opaque = opaque;
    return ts;
}

void qemu_free_timer(QEMUTimer *ts)
{
    qemu_free(ts);
}

/* stop a timer, but do not dealloc it */
void qemu_del_timer(QEMUTimer *ts)
{
    QEMUTimer **pt, *t;

    /* NOTE: this code must be signal safe because
       qemu_timer_expired() can be called from a signal. */
    pt = &active_timers[ts->clock->type];
    for(;;) {
        t = *pt;
        if (!t)
            break;
        if (t == ts) {
            *pt = t->next;
            break;
        }
        pt = &t->next;
    }
}

/* modify the current timer so that it will be fired when current_time
   >= expire_time. The corresponding callback will be called. */
void qemu_mod_timer(QEMUTimer *ts, int64_t expire_time)
{
    QEMUTimer **pt, *t;

    qemu_del_timer(ts);

    /* add the timer in the sorted list */
    /* NOTE: this code must be signal safe because
       qemu_timer_expired() can be called from a signal. */
    pt = &active_timers[ts->clock->type];
    for(;;) {
        t = *pt;
        if (!t)
            break;
902
        if (t->expire_time > expire_time)
903 904 905 906 907 908
            break;
        pt = &t->next;
    }
    ts->expire_time = expire_time;
    ts->next = *pt;
    *pt = ts;
909 910

    /* Rearm if necessary  */
P
pbrook 已提交
911
    if (pt == &active_timers[ts->clock->type]) {
P
Paolo Bonzini 已提交
912
        if (!alarm_timer->pending) {
P
pbrook 已提交
913 914 915
            qemu_rearm_alarm_timer(alarm_timer);
        }
        /* Interrupt execution to force deadline recalculation.  */
916 917
        if (use_icount)
            qemu_notify_event();
P
pbrook 已提交
918
    }
919 920 921 922 923 924 925 926 927 928 929 930
}

int qemu_timer_pending(QEMUTimer *ts)
{
    QEMUTimer *t;
    for(t = active_timers[ts->clock->type]; t != NULL; t = t->next) {
        if (t == ts)
            return 1;
    }
    return 0;
}

S
Stefano Stabellini 已提交
931
int qemu_timer_expired(QEMUTimer *timer_head, int64_t current_time)
932 933 934 935 936 937
{
    if (!timer_head)
        return 0;
    return (timer_head->expire_time <= current_time);
}

P
Paolo Bonzini 已提交
938
static void qemu_run_timers(QEMUClock *clock)
939
{
P
Paolo Bonzini 已提交
940 941
    QEMUTimer **ptimer_head, *ts;
    int64_t current_time;
942 943 944
   
    if (!clock->enabled)
        return;
945

P
Paolo Bonzini 已提交
946 947
    current_time = qemu_get_clock (clock);
    ptimer_head = &active_timers[clock->type];
948 949
    for(;;) {
        ts = *ptimer_head;
950
        if (!ts || ts->expire_time > current_time)
951 952 953 954
            break;
        /* remove timer from the list before calling the callback */
        *ptimer_head = ts->next;
        ts->next = NULL;
955

956 957 958 959 960 961 962 963
        /* run the callback (the timer list can be modified) */
        ts->cb(ts->opaque);
    }
}

int64_t qemu_get_clock(QEMUClock *clock)
{
    switch(clock->type) {
J
Jan Kiszka 已提交
964
    case QEMU_CLOCK_REALTIME:
965
        return get_clock() / 1000000;
966
    default:
J
Jan Kiszka 已提交
967
    case QEMU_CLOCK_VIRTUAL:
P
pbrook 已提交
968 969 970 971 972
        if (use_icount) {
            return cpu_get_icount();
        } else {
            return cpu_get_clock();
        }
J
Jan Kiszka 已提交
973 974
    case QEMU_CLOCK_HOST:
        return get_clock_realtime();
975 976 977
    }
}

P
Paolo Bonzini 已提交
978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994
int64_t qemu_get_clock_ns(QEMUClock *clock)
{
    switch(clock->type) {
    case QEMU_CLOCK_REALTIME:
        return get_clock();
    default:
    case QEMU_CLOCK_VIRTUAL:
        if (use_icount) {
            return cpu_get_icount();
        } else {
            return cpu_get_clock();
        }
    case QEMU_CLOCK_HOST:
        return get_clock_realtime();
    }
}

J
Jan Kiszka 已提交
995
static void init_clocks(void)
996 997
{
    init_get_clock();
J
Jan Kiszka 已提交
998 999
    rt_clock = qemu_new_clock(QEMU_CLOCK_REALTIME);
    vm_clock = qemu_new_clock(QEMU_CLOCK_VIRTUAL);
J
Jan Kiszka 已提交
1000
    host_clock = qemu_new_clock(QEMU_CLOCK_HOST);
J
Jan Kiszka 已提交
1001 1002

    rtc_clock = host_clock;
1003 1004
}

1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029
/* save a timer */
void qemu_put_timer(QEMUFile *f, QEMUTimer *ts)
{
    uint64_t expire_time;

    if (qemu_timer_pending(ts)) {
        expire_time = ts->expire_time;
    } else {
        expire_time = -1;
    }
    qemu_put_be64(f, expire_time);
}

void qemu_get_timer(QEMUFile *f, QEMUTimer *ts)
{
    uint64_t expire_time;

    expire_time = qemu_get_be64(f);
    if (expire_time != -1) {
        qemu_mod_timer(ts, expire_time);
    } else {
        qemu_del_timer(ts);
    }
}

J
Juan Quintela 已提交
1030 1031 1032 1033 1034 1035 1036
static const VMStateDescription vmstate_timers = {
    .name = "timer",
    .version_id = 2,
    .minimum_version_id = 1,
    .minimum_version_id_old = 1,
    .fields      = (VMStateField []) {
        VMSTATE_INT64(cpu_ticks_offset, TimersState),
1037
        VMSTATE_INT64(dummy, TimersState),
J
Juan Quintela 已提交
1038 1039
        VMSTATE_INT64_V(cpu_clock_offset, TimersState, 2),
        VMSTATE_END_OF_LIST()
1040
    }
J
Juan Quintela 已提交
1041
};
1042

1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
static void qemu_run_all_timers(void)
{
    /* rearm timer, if not periodic */
    if (alarm_timer->expired) {
        alarm_timer->expired = 0;
        qemu_rearm_alarm_timer(alarm_timer);
    }

    alarm_timer->pending = 0;

    /* vm time timers */
    if (vm_running) {
1055
        qemu_run_timers(vm_clock);
1056 1057
    }

P
Paolo Bonzini 已提交
1058 1059
    qemu_run_timers(rt_clock);
    qemu_run_timers(host_clock);
1060
}
1061

B
bellard 已提交
1062
#ifdef _WIN32
B
blueswir1 已提交
1063 1064 1065
static void CALLBACK host_alarm_handler(UINT uTimerID, UINT uMsg,
                                        DWORD_PTR dwUser, DWORD_PTR dw1,
                                        DWORD_PTR dw2)
B
bellard 已提交
1066
#else
1067
static void host_alarm_handler(int host_signum)
B
bellard 已提交
1068
#endif
1069
{
P
Paolo Bonzini 已提交
1070 1071 1072 1073
    struct qemu_alarm_timer *t = alarm_timer;
    if (!t)
	return;

1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
#if 0
#define DISP_FREQ 1000
    {
        static int64_t delta_min = INT64_MAX;
        static int64_t delta_max, delta_cum, last_clock, delta, ti;
        static int count;
        ti = qemu_get_clock(vm_clock);
        if (last_clock != 0) {
            delta = ti - last_clock;
            if (delta < delta_min)
                delta_min = delta;
            if (delta > delta_max)
                delta_max = delta;
            delta_cum += delta;
            if (++count == DISP_FREQ) {
B
bellard 已提交
1089
                printf("timer: min=%" PRId64 " us max=%" PRId64 " us avg=%" PRId64 " us avg_freq=%0.3f Hz\n",
1090 1091 1092 1093
                       muldiv64(delta_min, 1000000, get_ticks_per_sec()),
                       muldiv64(delta_max, 1000000, get_ticks_per_sec()),
                       muldiv64(delta_cum, 1000000 / DISP_FREQ, get_ticks_per_sec()),
                       (double)get_ticks_per_sec() / ((double)delta_cum / DISP_FREQ));
1094 1095 1096 1097 1098 1099 1100 1101 1102
                count = 0;
                delta_min = INT64_MAX;
                delta_max = 0;
                delta_cum = 0;
            }
        }
        last_clock = ti;
    }
#endif
P
Paolo Bonzini 已提交
1103
    if (alarm_has_dynticks(t) ||
P
pbrook 已提交
1104
        (!use_icount &&
J
Jan Kiszka 已提交
1105
            qemu_timer_expired(active_timers[QEMU_CLOCK_VIRTUAL],
P
pbrook 已提交
1106
                               qemu_get_clock(vm_clock))) ||
J
Jan Kiszka 已提交
1107
        qemu_timer_expired(active_timers[QEMU_CLOCK_REALTIME],
J
Jan Kiszka 已提交
1108 1109 1110
                           qemu_get_clock(rt_clock)) ||
        qemu_timer_expired(active_timers[QEMU_CLOCK_HOST],
                           qemu_get_clock(host_clock))) {
1111

P
Paolo Bonzini 已提交
1112 1113
        t->expired = alarm_has_dynticks(t);
        t->pending = 1;
1114
        qemu_notify_event();
1115 1116 1117
    }
}

P
pbrook 已提交
1118
static int64_t qemu_next_deadline(void)
1119
{
J
Jan Kiszka 已提交
1120 1121
    /* To avoid problems with overflow limit this to 2^32.  */
    int64_t delta = INT32_MAX;
1122

J
Jan Kiszka 已提交
1123 1124
    if (active_timers[QEMU_CLOCK_VIRTUAL]) {
        delta = active_timers[QEMU_CLOCK_VIRTUAL]->expire_time -
P
pbrook 已提交
1125
                     qemu_get_clock(vm_clock);
J
Jan Kiszka 已提交
1126 1127 1128 1129 1130 1131
    }
    if (active_timers[QEMU_CLOCK_HOST]) {
        int64_t hdelta = active_timers[QEMU_CLOCK_HOST]->expire_time -
                 qemu_get_clock(host_clock);
        if (hdelta < delta)
            delta = hdelta;
1132 1133
    }

P
pbrook 已提交
1134 1135
    if (delta < 0)
        delta = 0;
1136

P
pbrook 已提交
1137 1138 1139
    return delta;
}

J
Jan Kiszka 已提交
1140
#if defined(__linux__)
P
pbrook 已提交
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
static uint64_t qemu_next_deadline_dyntick(void)
{
    int64_t delta;
    int64_t rtdelta;

    if (use_icount)
        delta = INT32_MAX;
    else
        delta = (qemu_next_deadline() + 999) / 1000;

J
Jan Kiszka 已提交
1151 1152
    if (active_timers[QEMU_CLOCK_REALTIME]) {
        rtdelta = (active_timers[QEMU_CLOCK_REALTIME]->expire_time -
P
pbrook 已提交
1153 1154 1155 1156 1157 1158 1159 1160 1161
                 qemu_get_clock(rt_clock))*1000;
        if (rtdelta < delta)
            delta = rtdelta;
    }

    if (delta < MIN_TIMER_REARM_US)
        delta = MIN_TIMER_REARM_US;

    return delta;
1162
}
1163
#endif
1164

1165 1166
#ifndef _WIN32

1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
/* Sets a specific flag */
static int fcntl_setfl(int fd, int flag)
{
    int flags;

    flags = fcntl(fd, F_GETFL);
    if (flags == -1)
        return -errno;

    if (fcntl(fd, F_SETFL, flags | flag) == -1)
        return -errno;

    return 0;
}

B
BSD fix  
bellard 已提交
1182 1183
#if defined(__linux__)

1184 1185
#define RTC_FREQ 1024

1186
static void enable_sigio_timer(int fd)
1187 1188 1189 1190 1191 1192 1193 1194 1195
{
    struct sigaction act;

    /* timer signal */
    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGIO, &act, NULL);
1196
    fcntl_setfl(fd, O_ASYNC);
1197 1198
    fcntl(fd, F_SETOWN, getpid());
}
B
BSD fix  
bellard 已提交
1199

T
ths 已提交
1200 1201 1202 1203 1204
static int hpet_start_timer(struct qemu_alarm_timer *t)
{
    struct hpet_info info;
    int r, fd;

K
Kevin Wolf 已提交
1205
    fd = qemu_open("/dev/hpet", O_RDONLY);
T
ths 已提交
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
    if (fd < 0)
        return -1;

    /* Set frequency */
    r = ioctl(fd, HPET_IRQFREQ, RTC_FREQ);
    if (r < 0) {
        fprintf(stderr, "Could not configure '/dev/hpet' to have a 1024Hz timer. This is not a fatal\n"
                "error, but for better emulation accuracy type:\n"
                "'echo 1024 > /proc/sys/dev/hpet/max-user-freq' as root.\n");
        goto fail;
    }

    /* Check capabilities */
    r = ioctl(fd, HPET_INFO, &info);
    if (r < 0)
        goto fail;

    /* Enable periodic mode */
    r = ioctl(fd, HPET_EPI, 0);
    if (info.hi_flags && (r < 0))
        goto fail;

    /* Enable interrupt */
    r = ioctl(fd, HPET_IE_ON, 0);
    if (r < 0)
        goto fail;

    enable_sigio_timer(fd);
1234
    t->priv = (void *)(long)fd;
T
ths 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243

    return 0;
fail:
    close(fd);
    return -1;
}

static void hpet_stop_timer(struct qemu_alarm_timer *t)
{
1244
    int fd = (long)t->priv;
T
ths 已提交
1245 1246 1247 1248

    close(fd);
}

1249
static int rtc_start_timer(struct qemu_alarm_timer *t)
1250
{
1251
    int rtc_fd;
1252
    unsigned long current_rtc_freq = 0;
1253

K
Kevin Wolf 已提交
1254
    TFR(rtc_fd = qemu_open("/dev/rtc", O_RDONLY));
1255 1256
    if (rtc_fd < 0)
        return -1;
1257 1258 1259
    ioctl(rtc_fd, RTC_IRQP_READ, &current_rtc_freq);
    if (current_rtc_freq != RTC_FREQ &&
        ioctl(rtc_fd, RTC_IRQP_SET, RTC_FREQ) < 0) {
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
        fprintf(stderr, "Could not configure '/dev/rtc' to have a 1024 Hz timer. This is not a fatal\n"
                "error, but for better emulation accuracy either use a 2.6 host Linux kernel or\n"
                "type 'echo 1024 > /proc/sys/dev/rtc/max-user-freq' as root.\n");
        goto fail;
    }
    if (ioctl(rtc_fd, RTC_PIE_ON, 0) < 0) {
    fail:
        close(rtc_fd);
        return -1;
    }
1270 1271 1272

    enable_sigio_timer(rtc_fd);

1273
    t->priv = (void *)(long)rtc_fd;
1274

1275 1276 1277
    return 0;
}

1278
static void rtc_stop_timer(struct qemu_alarm_timer *t)
B
BSD fix  
bellard 已提交
1279
{
1280
    int rtc_fd = (long)t->priv;
1281 1282

    close(rtc_fd);
B
BSD fix  
bellard 已提交
1283 1284
}

1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
static int dynticks_start_timer(struct qemu_alarm_timer *t)
{
    struct sigevent ev;
    timer_t host_timer;
    struct sigaction act;

    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGALRM, &act, NULL);

1297 1298 1299 1300 1301
    /* 
     * Initialize ev struct to 0 to avoid valgrind complaining
     * about uninitialized data in timer_create call
     */
    memset(&ev, 0, sizeof(ev));
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
    ev.sigev_value.sival_int = 0;
    ev.sigev_notify = SIGEV_SIGNAL;
    ev.sigev_signo = SIGALRM;

    if (timer_create(CLOCK_REALTIME, &ev, &host_timer)) {
        perror("timer_create");

        /* disable dynticks */
        fprintf(stderr, "Dynamic Ticks disabled\n");

        return -1;
    }

1315
    t->priv = (void *)(long)host_timer;
1316 1317 1318 1319 1320 1321

    return 0;
}

static void dynticks_stop_timer(struct qemu_alarm_timer *t)
{
1322
    timer_t host_timer = (timer_t)(long)t->priv;
1323 1324 1325 1326 1327 1328

    timer_delete(host_timer);
}

static void dynticks_rearm_timer(struct qemu_alarm_timer *t)
{
1329
    timer_t host_timer = (timer_t)(long)t->priv;
1330 1331 1332 1333
    struct itimerspec timeout;
    int64_t nearest_delta_us = INT64_MAX;
    int64_t current_us;

1334
    assert(alarm_has_dynticks(t));
J
Jan Kiszka 已提交
1335
    if (!active_timers[QEMU_CLOCK_REALTIME] &&
J
Jan Kiszka 已提交
1336 1337
        !active_timers[QEMU_CLOCK_VIRTUAL] &&
        !active_timers[QEMU_CLOCK_HOST])
1338
        return;
1339

P
pbrook 已提交
1340
    nearest_delta_us = qemu_next_deadline_dyntick();
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362

    /* check whether a timer is already running */
    if (timer_gettime(host_timer, &timeout)) {
        perror("gettime");
        fprintf(stderr, "Internal timer error: aborting\n");
        exit(1);
    }
    current_us = timeout.it_value.tv_sec * 1000000 + timeout.it_value.tv_nsec/1000;
    if (current_us && current_us <= nearest_delta_us)
        return;

    timeout.it_interval.tv_sec = 0;
    timeout.it_interval.tv_nsec = 0; /* 0 for one-shot timer */
    timeout.it_value.tv_sec =  nearest_delta_us / 1000000;
    timeout.it_value.tv_nsec = (nearest_delta_us % 1000000) * 1000;
    if (timer_settime(host_timer, 0 /* RELATIVE */, &timeout, NULL)) {
        perror("settime");
        fprintf(stderr, "Internal timer error: aborting\n");
        exit(1);
    }
}

T
ths 已提交
1363
#endif /* defined(__linux__) */
1364

1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
static int unix_start_timer(struct qemu_alarm_timer *t)
{
    struct sigaction act;
    struct itimerval itv;
    int err;

    /* timer signal */
    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGALRM, &act, NULL);

    itv.it_interval.tv_sec = 0;
    /* for i386 kernel 2.6 to get 1 ms */
    itv.it_interval.tv_usec = 999;
    itv.it_value.tv_sec = 0;
    itv.it_value.tv_usec = 10 * 1000;

    err = setitimer(ITIMER_REAL, &itv, NULL);
    if (err)
        return -1;

    return 0;
}

static void unix_stop_timer(struct qemu_alarm_timer *t)
{
    struct itimerval itv;

    memset(&itv, 0, sizeof(itv));
    setitimer(ITIMER_REAL, &itv, NULL);
}

B
BSD fix  
bellard 已提交
1399
#endif /* !defined(_WIN32) */
1400

1401

1402 1403 1404 1405 1406 1407
#ifdef _WIN32

static int win32_start_timer(struct qemu_alarm_timer *t)
{
    TIMECAPS tc;
    struct qemu_alarm_win32 *data = t->priv;
1408
    UINT flags;
1409 1410 1411 1412

    memset(&tc, 0, sizeof(tc));
    timeGetDevCaps(&tc, sizeof(tc));

1413
    data->period = tc.wPeriodMin;
1414 1415
    timeBeginPeriod(data->period);

1416 1417 1418 1419 1420 1421
    flags = TIME_CALLBACK_FUNCTION;
    if (alarm_has_dynticks(t))
        flags |= TIME_ONESHOT;
    else
        flags |= TIME_PERIODIC;

1422 1423 1424 1425
    data->timerId = timeSetEvent(1,         // interval (ms)
                        data->period,       // resolution
                        host_alarm_handler, // function
                        (DWORD)t,           // parameter
1426
                        flags);
1427 1428

    if (!data->timerId) {
B
Blue Swirl 已提交
1429
        fprintf(stderr, "Failed to initialize win32 alarm timer: %ld\n",
1430
                GetLastError());
1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
        timeEndPeriod(data->period);
        return -1;
    }

    return 0;
}

static void win32_stop_timer(struct qemu_alarm_timer *t)
{
    struct qemu_alarm_win32 *data = t->priv;

    timeKillEvent(data->timerId);
    timeEndPeriod(data->period);
}

1446 1447 1448 1449
static void win32_rearm_timer(struct qemu_alarm_timer *t)
{
    struct qemu_alarm_win32 *data = t->priv;

1450
    assert(alarm_has_dynticks(t));
J
Jan Kiszka 已提交
1451
    if (!active_timers[QEMU_CLOCK_REALTIME] &&
J
Jan Kiszka 已提交
1452 1453
        !active_timers[QEMU_CLOCK_VIRTUAL] &&
        !active_timers[QEMU_CLOCK_HOST])
1454
        return;
1455 1456 1457 1458 1459 1460 1461

    timeKillEvent(data->timerId);

    data->timerId = timeSetEvent(1,
                        data->period,
                        host_alarm_handler,
                        (DWORD)t,
P
Paolo Bonzini 已提交
1462
                        TIME_ONESHOT | TIME_CALLBACK_FUNCTION);
1463 1464

    if (!data->timerId) {
B
Blue Swirl 已提交
1465
        fprintf(stderr, "Failed to re-arm win32 alarm timer %ld\n",
1466
                GetLastError());
1467 1468 1469 1470 1471 1472

        timeEndPeriod(data->period);
        exit(1);
    }
}

1473 1474
#endif /* _WIN32 */

1475 1476 1477 1478 1479 1480
static void alarm_timer_on_change_state_rearm(void *opaque, int running, int reason)
{
    if (running)
        qemu_rearm_alarm_timer((struct qemu_alarm_timer *) opaque);
}

1481
static int init_timer_alarm(void)
1482
{
1483
    struct qemu_alarm_timer *t = NULL;
1484
    int i, err = -1;
1485

1486 1487 1488 1489 1490 1491
    for (i = 0; alarm_timers[i].name; i++) {
        t = &alarm_timers[i];

        err = t->start(t);
        if (!err)
            break;
B
bellard 已提交
1492
    }
1493

1494
    if (err) {
1495 1496
        err = -ENOENT;
        goto fail;
B
bellard 已提交
1497
    }
1498

P
Paolo Bonzini 已提交
1499 1500
    /* first event is at time 0 */
    t->pending = 1;
1501
    alarm_timer = t;
1502
    qemu_add_vm_change_state_handler(alarm_timer_on_change_state_rearm, t);
1503

1504
    return 0;
1505 1506 1507

fail:
    return err;
1508 1509
}

1510
static void quit_timers(void)
B
bellard 已提交
1511
{
P
Paolo Bonzini 已提交
1512
    struct qemu_alarm_timer *t = alarm_timer;
1513
    alarm_timer = NULL;
P
Paolo Bonzini 已提交
1514
    t->stop(t);
B
bellard 已提交
1515 1516
}

1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
/***********************************************************/
/* host time/date access */
void qemu_get_timedate(struct tm *tm, int offset)
{
    time_t ti;
    struct tm *ret;

    time(&ti);
    ti += offset;
    if (rtc_date_offset == -1) {
        if (rtc_utc)
            ret = gmtime(&ti);
        else
            ret = localtime(&ti);
    } else {
        ti -= rtc_date_offset;
        ret = gmtime(&ti);
    }

    memcpy(tm, ret, sizeof(struct tm));
}

int qemu_timedate_diff(struct tm *tm)
{
    time_t seconds;

    if (rtc_date_offset == -1)
        if (rtc_utc)
            seconds = mktimegm(tm);
        else
            seconds = mktime(tm);
    else
        seconds = mktimegm(tm) + rtc_date_offset;

    return seconds - time(NULL);
}

1554 1555 1556 1557 1558 1559 1560 1561 1562
void rtc_change_mon_event(struct tm *tm)
{
    QObject *data;

    data = qobject_from_jsonf("{ 'offset': %d }", qemu_timedate_diff(tm));
    monitor_protocol_event(QEVENT_RTC_CHANGE, data);
    qobject_decref(data);
}

J
Jan Kiszka 已提交
1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
static void configure_rtc_date_offset(const char *startdate, int legacy)
{
    time_t rtc_start_date;
    struct tm tm;

    if (!strcmp(startdate, "now") && legacy) {
        rtc_date_offset = -1;
    } else {
        if (sscanf(startdate, "%d-%d-%dT%d:%d:%d",
                   &tm.tm_year,
                   &tm.tm_mon,
                   &tm.tm_mday,
                   &tm.tm_hour,
                   &tm.tm_min,
                   &tm.tm_sec) == 6) {
            /* OK */
        } else if (sscanf(startdate, "%d-%d-%d",
                          &tm.tm_year,
                          &tm.tm_mon,
                          &tm.tm_mday) == 3) {
            tm.tm_hour = 0;
            tm.tm_min = 0;
            tm.tm_sec = 0;
        } else {
            goto date_fail;
        }
        tm.tm_year -= 1900;
        tm.tm_mon--;
        rtc_start_date = mktimegm(&tm);
        if (rtc_start_date == -1) {
        date_fail:
            fprintf(stderr, "Invalid date format. Valid formats are:\n"
                            "'2006-06-17T16:01:21' or '2006-06-17'\n");
            exit(1);
        }
        rtc_date_offset = time(NULL) - rtc_start_date;
    }
}

static void configure_rtc(QemuOpts *opts)
{
    const char *value;

    value = qemu_opt_get(opts, "base");
    if (value) {
        if (!strcmp(value, "utc")) {
            rtc_utc = 1;
        } else if (!strcmp(value, "localtime")) {
            rtc_utc = 0;
        } else {
            configure_rtc_date_offset(value, 0);
        }
    }
J
Jan Kiszka 已提交
1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626
    value = qemu_opt_get(opts, "clock");
    if (value) {
        if (!strcmp(value, "host")) {
            rtc_clock = host_clock;
        } else if (!strcmp(value, "vm")) {
            rtc_clock = vm_clock;
        } else {
            fprintf(stderr, "qemu: invalid option value '%s'\n", value);
            exit(1);
        }
    }
J
Jan Kiszka 已提交
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
#ifdef CONFIG_TARGET_I386
    value = qemu_opt_get(opts, "driftfix");
    if (value) {
        if (!strcmp(buf, "slew")) {
            rtc_td_hack = 1;
        } else if (!strcmp(buf, "none")) {
            rtc_td_hack = 0;
        } else {
            fprintf(stderr, "qemu: invalid option value '%s'\n", value);
            exit(1);
        }
    }
#endif
}

B
bellard 已提交
1642 1643 1644 1645 1646
#ifdef _WIN32
static void socket_cleanup(void)
{
    WSACleanup();
}
B
bellard 已提交
1647

B
bellard 已提交
1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661
static int socket_init(void)
{
    WSADATA Data;
    int ret, err;

    ret = WSAStartup(MAKEWORD(2,2), &Data);
    if (ret != 0) {
        err = WSAGetLastError();
        fprintf(stderr, "WSAStartup: %d\n", err);
        return -1;
    }
    atexit(socket_cleanup);
    return 0;
}
1662 1663
#endif

1664 1665 1666 1667 1668
/***********************************************************/
/* Bluetooth support */
static int nb_hcis;
static int cur_hci;
static struct HCIInfo *hci_table[MAX_NICS];
1669

1670 1671 1672 1673 1674 1675 1676
static struct bt_vlan_s {
    struct bt_scatternet_s net;
    int id;
    struct bt_vlan_s *next;
} *first_bt_vlan;

/* find or alloc a new bluetooth "VLAN" */
B
blueswir1 已提交
1677
static struct bt_scatternet_s *qemu_find_bt_vlan(int id)
1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716
{
    struct bt_vlan_s **pvlan, *vlan;
    for (vlan = first_bt_vlan; vlan != NULL; vlan = vlan->next) {
        if (vlan->id == id)
            return &vlan->net;
    }
    vlan = qemu_mallocz(sizeof(struct bt_vlan_s));
    vlan->id = id;
    pvlan = &first_bt_vlan;
    while (*pvlan != NULL)
        pvlan = &(*pvlan)->next;
    *pvlan = vlan;
    return &vlan->net;
}

static void null_hci_send(struct HCIInfo *hci, const uint8_t *data, int len)
{
}

static int null_hci_addr_set(struct HCIInfo *hci, const uint8_t *bd_addr)
{
    return -ENOTSUP;
}

static struct HCIInfo null_hci = {
    .cmd_send = null_hci_send,
    .sco_send = null_hci_send,
    .acl_send = null_hci_send,
    .bdaddr_set = null_hci_addr_set,
};

struct HCIInfo *qemu_next_hci(void)
{
    if (cur_hci == nb_hcis)
        return &null_hci;

    return hci_table[cur_hci++];
}

1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749
static struct HCIInfo *hci_init(const char *str)
{
    char *endp;
    struct bt_scatternet_s *vlan = 0;

    if (!strcmp(str, "null"))
        /* null */
        return &null_hci;
    else if (!strncmp(str, "host", 4) && (str[4] == '\0' || str[4] == ':'))
        /* host[:hciN] */
        return bt_host_hci(str[4] ? str + 5 : "hci0");
    else if (!strncmp(str, "hci", 3)) {
        /* hci[,vlan=n] */
        if (str[3]) {
            if (!strncmp(str + 3, ",vlan=", 6)) {
                vlan = qemu_find_bt_vlan(strtol(str + 9, &endp, 0));
                if (*endp)
                    vlan = 0;
            }
        } else
            vlan = qemu_find_bt_vlan(0);
        if (vlan)
           return bt_new_hci(vlan);
    }

    fprintf(stderr, "qemu: Unknown bluetooth HCI `%s'.\n", str);

    return 0;
}

static int bt_hci_parse(const char *str)
{
    struct HCIInfo *hci;
A
Anthony Liguori 已提交
1750
    bdaddr_t bdaddr;
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854

    if (nb_hcis >= MAX_NICS) {
        fprintf(stderr, "qemu: Too many bluetooth HCIs (max %i).\n", MAX_NICS);
        return -1;
    }

    hci = hci_init(str);
    if (!hci)
        return -1;

    bdaddr.b[0] = 0x52;
    bdaddr.b[1] = 0x54;
    bdaddr.b[2] = 0x00;
    bdaddr.b[3] = 0x12;
    bdaddr.b[4] = 0x34;
    bdaddr.b[5] = 0x56 + nb_hcis;
    hci->bdaddr_set(hci, bdaddr.b);

    hci_table[nb_hcis++] = hci;

    return 0;
}

static void bt_vhci_add(int vlan_id)
{
    struct bt_scatternet_s *vlan = qemu_find_bt_vlan(vlan_id);

    if (!vlan->slave)
        fprintf(stderr, "qemu: warning: adding a VHCI to "
                        "an empty scatternet %i\n", vlan_id);

    bt_vhci_init(bt_new_hci(vlan));
}

static struct bt_device_s *bt_device_add(const char *opt)
{
    struct bt_scatternet_s *vlan;
    int vlan_id = 0;
    char *endp = strstr(opt, ",vlan=");
    int len = (endp ? endp - opt : strlen(opt)) + 1;
    char devname[10];

    pstrcpy(devname, MIN(sizeof(devname), len), opt);

    if (endp) {
        vlan_id = strtol(endp + 6, &endp, 0);
        if (*endp) {
            fprintf(stderr, "qemu: unrecognised bluetooth vlan Id\n");
            return 0;
        }
    }

    vlan = qemu_find_bt_vlan(vlan_id);

    if (!vlan->slave)
        fprintf(stderr, "qemu: warning: adding a slave device to "
                        "an empty scatternet %i\n", vlan_id);

    if (!strcmp(devname, "keyboard"))
        return bt_keyboard_init(vlan);

    fprintf(stderr, "qemu: unsupported bluetooth device `%s'\n", devname);
    return 0;
}

static int bt_parse(const char *opt)
{
    const char *endp, *p;
    int vlan;

    if (strstart(opt, "hci", &endp)) {
        if (!*endp || *endp == ',') {
            if (*endp)
                if (!strstart(endp, ",vlan=", 0))
                    opt = endp + 1;

            return bt_hci_parse(opt);
       }
    } else if (strstart(opt, "vhci", &endp)) {
        if (!*endp || *endp == ',') {
            if (*endp) {
                if (strstart(endp, ",vlan=", &p)) {
                    vlan = strtol(p, (char **) &endp, 0);
                    if (*endp) {
                        fprintf(stderr, "qemu: bad scatternet '%s'\n", p);
                        return 1;
                    }
                } else {
                    fprintf(stderr, "qemu: bad parameter '%s'\n", endp + 1);
                    return 1;
                }
            } else
                vlan = 0;

            bt_vhci_add(vlan);
            return 0;
        }
    } else if (strstart(opt, "device:", &endp))
        return !bt_device_add(endp);

    fprintf(stderr, "qemu: bad bluetooth parameter '%s'\n", opt);
    return 1;
}

1855 1856 1857
/***********************************************************/
/* QEMU Block devices */

1858
#define HD_ALIAS "index=%d,media=disk"
T
ths 已提交
1859 1860
#define CDROM_ALIAS "index=2,media=cdrom"
#define FD_ALIAS "index=%d,if=floppy"
1861 1862
#define PFLASH_ALIAS "if=pflash"
#define MTD_ALIAS "if=mtd"
1863
#define SD_ALIAS "index=0,if=sd"
T
ths 已提交
1864

G
Gerd Hoffmann 已提交
1865
QemuOpts *drive_add(const char *file, const char *fmt, ...)
T
ths 已提交
1866 1867
{
    va_list ap;
G
Gerd Hoffmann 已提交
1868 1869
    char optstr[1024];
    QemuOpts *opts;
T
ths 已提交
1870 1871

    va_start(ap, fmt);
G
Gerd Hoffmann 已提交
1872
    vsnprintf(optstr, sizeof(optstr), fmt, ap);
T
ths 已提交
1873 1874
    va_end(ap);

1875
    opts = qemu_opts_parse(&qemu_drive_opts, optstr, 0);
G
Gerd Hoffmann 已提交
1876 1877 1878 1879 1880 1881 1882 1883
    if (!opts) {
        fprintf(stderr, "%s: huh? duplicate? (%s)\n",
                __FUNCTION__, optstr);
        return NULL;
    }
    if (file)
        qemu_opt_set(opts, "file", file);
    return opts;
1884 1885
}

G
Gerd Hoffmann 已提交
1886
DriveInfo *drive_get(BlockInterfaceType type, int bus, int unit)
T
ths 已提交
1887
{
G
Gerd Hoffmann 已提交
1888
    DriveInfo *dinfo;
T
ths 已提交
1889 1890 1891

    /* seek interface, bus and unit */

B
Blue Swirl 已提交
1892
    QTAILQ_FOREACH(dinfo, &drives, next) {
G
Gerd Hoffmann 已提交
1893 1894 1895 1896 1897
        if (dinfo->type == type &&
	    dinfo->bus == bus &&
	    dinfo->unit == unit)
            return dinfo;
    }
T
ths 已提交
1898

G
Gerd Hoffmann 已提交
1899
    return NULL;
T
ths 已提交
1900 1901
}

G
Gerd Hoffmann 已提交
1902
DriveInfo *drive_get_by_id(const char *id)
G
Gerd Hoffmann 已提交
1903 1904 1905
{
    DriveInfo *dinfo;

B
Blue Swirl 已提交
1906
    QTAILQ_FOREACH(dinfo, &drives, next) {
G
Gerd Hoffmann 已提交
1907 1908 1909 1910 1911 1912 1913
        if (strcmp(id, dinfo->id))
            continue;
        return dinfo;
    }
    return NULL;
}

1914
int drive_get_max_bus(BlockInterfaceType type)
T
ths 已提交
1915 1916
{
    int max_bus;
G
Gerd Hoffmann 已提交
1917
    DriveInfo *dinfo;
T
ths 已提交
1918 1919

    max_bus = -1;
B
Blue Swirl 已提交
1920
    QTAILQ_FOREACH(dinfo, &drives, next) {
G
Gerd Hoffmann 已提交
1921 1922 1923
        if(dinfo->type == type &&
           dinfo->bus > max_bus)
            max_bus = dinfo->bus;
T
ths 已提交
1924 1925 1926 1927
    }
    return max_bus;
}

1928 1929
const char *drive_get_serial(BlockDriverState *bdrv)
{
G
Gerd Hoffmann 已提交
1930
    DriveInfo *dinfo;
1931

B
Blue Swirl 已提交
1932
    QTAILQ_FOREACH(dinfo, &drives, next) {
G
Gerd Hoffmann 已提交
1933 1934 1935
        if (dinfo->bdrv == bdrv)
            return dinfo->serial;
    }
1936 1937 1938 1939

    return "\0";
}

1940 1941
BlockInterfaceErrorAction drive_get_on_error(
    BlockDriverState *bdrv, int is_read)
1942
{
G
Gerd Hoffmann 已提交
1943
    DriveInfo *dinfo;
1944

B
Blue Swirl 已提交
1945
    QTAILQ_FOREACH(dinfo, &drives, next) {
G
Gerd Hoffmann 已提交
1946
        if (dinfo->bdrv == bdrv)
K
Kevin Wolf 已提交
1947
            return is_read ? dinfo->on_read_error : dinfo->on_write_error;
G
Gerd Hoffmann 已提交
1948
    }
1949

K
Kevin Wolf 已提交
1950
    return is_read ? BLOCK_ERR_REPORT : BLOCK_ERR_STOP_ENOSPC;
1951 1952
}

1953 1954 1955 1956 1957
static void bdrv_format_print(void *opaque, const char *name)
{
    fprintf(stderr, " %s", name);
}

G
Gerd Hoffmann 已提交
1958
void drive_uninit(DriveInfo *dinfo)
1959
{
G
Gerd Hoffmann 已提交
1960 1961 1962 1963
    qemu_opts_del(dinfo->opts);
    bdrv_delete(dinfo->bdrv);
    QTAILQ_REMOVE(&drives, dinfo, next);
    qemu_free(dinfo);
1964 1965
}

K
Kevin Wolf 已提交
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
static int parse_block_error_action(const char *buf, int is_read)
{
    if (!strcmp(buf, "ignore")) {
        return BLOCK_ERR_IGNORE;
    } else if (!is_read && !strcmp(buf, "enospc")) {
        return BLOCK_ERR_STOP_ENOSPC;
    } else if (!strcmp(buf, "stop")) {
        return BLOCK_ERR_STOP_ANY;
    } else if (!strcmp(buf, "report")) {
        return BLOCK_ERR_REPORT;
    } else {
        fprintf(stderr, "qemu: '%s' invalid %s error action\n",
            buf, is_read ? "read" : "write");
        return -1;
    }
}

G
Gerd Hoffmann 已提交
1983
DriveInfo *drive_init(QemuOpts *opts, void *opaque,
G
Gerd Hoffmann 已提交
1984
                      int *fatal_error)
T
ths 已提交
1985
{
G
Gerd Hoffmann 已提交
1986 1987
    const char *buf;
    const char *file = NULL;
1988
    char devname[128];
G
Gerd Hoffmann 已提交
1989
    const char *serial;
1990
    const char *mediastr = "";
1991
    BlockInterfaceType type;
T
ths 已提交
1992 1993 1994
    enum { MEDIA_DISK, MEDIA_CDROM } media;
    int bus_id, unit_id;
    int cyls, heads, secs, translation;
1995
    BlockDriver *drv = NULL;
1996
    QEMUMachine *machine = opaque;
T
ths 已提交
1997 1998
    int max_devs;
    int index;
1999
    int cache;
2000
    int aio = 0;
2001
    int ro = 0;
K
Kevin Wolf 已提交
2002 2003
    int bdrv_flags;
    int on_read_error, on_write_error;
2004
    const char *devaddr;
G
Gerd Hoffmann 已提交
2005
    DriveInfo *dinfo;
G
Gerd Hoffmann 已提交
2006
    int snapshot = 0;
T
ths 已提交
2007

G
Gerd Hoffmann 已提交
2008
    *fatal_error = 1;
T
ths 已提交
2009 2010

    translation = BIOS_ATA_TRANSLATION_AUTO;
2011
    cache = 1;
T
ths 已提交
2012

2013
    if (machine && machine->use_scsi) {
2014
        type = IF_SCSI;
T
ths 已提交
2015
        max_devs = MAX_SCSI_DEVS;
B
blueswir1 已提交
2016
        pstrcpy(devname, sizeof(devname), "scsi");
T
ths 已提交
2017
    } else {
2018
        type = IF_IDE;
T
ths 已提交
2019
        max_devs = MAX_IDE_DEVS;
B
blueswir1 已提交
2020
        pstrcpy(devname, sizeof(devname), "ide");
T
ths 已提交
2021 2022 2023 2024
    }
    media = MEDIA_DISK;

    /* extract parameters */
G
Gerd Hoffmann 已提交
2025 2026 2027
    bus_id  = qemu_opt_get_number(opts, "bus", 0);
    unit_id = qemu_opt_get_number(opts, "unit", -1);
    index   = qemu_opt_get_number(opts, "index", -1);
T
ths 已提交
2028

G
Gerd Hoffmann 已提交
2029 2030 2031
    cyls  = qemu_opt_get_number(opts, "cyls", 0);
    heads = qemu_opt_get_number(opts, "heads", 0);
    secs  = qemu_opt_get_number(opts, "secs", 0);
T
ths 已提交
2032

G
Gerd Hoffmann 已提交
2033
    snapshot = qemu_opt_get_bool(opts, "snapshot", 0);
2034
    ro = qemu_opt_get_bool(opts, "readonly", 0);
T
ths 已提交
2035

G
Gerd Hoffmann 已提交
2036 2037 2038 2039
    file = qemu_opt_get(opts, "file");
    serial = qemu_opt_get(opts, "serial");

    if ((buf = qemu_opt_get(opts, "if")) != NULL) {
B
bellard 已提交
2040
        pstrcpy(devname, sizeof(devname), buf);
T
ths 已提交
2041
        if (!strcmp(buf, "ide")) {
2042
	    type = IF_IDE;
T
ths 已提交
2043 2044
            max_devs = MAX_IDE_DEVS;
        } else if (!strcmp(buf, "scsi")) {
2045
	    type = IF_SCSI;
T
ths 已提交
2046 2047
            max_devs = MAX_SCSI_DEVS;
        } else if (!strcmp(buf, "floppy")) {
2048
	    type = IF_FLOPPY;
T
ths 已提交
2049 2050
            max_devs = 0;
        } else if (!strcmp(buf, "pflash")) {
2051
	    type = IF_PFLASH;
T
ths 已提交
2052 2053
            max_devs = 0;
	} else if (!strcmp(buf, "mtd")) {
2054
	    type = IF_MTD;
T
ths 已提交
2055 2056
            max_devs = 0;
	} else if (!strcmp(buf, "sd")) {
2057
	    type = IF_SD;
T
ths 已提交
2058
            max_devs = 0;
A
aliguori 已提交
2059 2060 2061
        } else if (!strcmp(buf, "virtio")) {
            type = IF_VIRTIO;
            max_devs = 0;
2062 2063 2064
	} else if (!strcmp(buf, "xen")) {
	    type = IF_XEN;
            max_devs = 0;
G
Gerd Hoffmann 已提交
2065 2066 2067
	} else if (!strcmp(buf, "none")) {
	    type = IF_NONE;
            max_devs = 0;
2068
	} else {
G
Gerd Hoffmann 已提交
2069
            fprintf(stderr, "qemu: unsupported bus type '%s'\n", buf);
G
Gerd Hoffmann 已提交
2070
            return NULL;
T
ths 已提交
2071 2072 2073 2074
	}
    }

    if (cyls || heads || secs) {
2075
        if (cyls < 1 || (type == IF_IDE && cyls > 16383)) {
G
Gerd Hoffmann 已提交
2076
            fprintf(stderr, "qemu: '%s' invalid physical cyls number\n", buf);
G
Gerd Hoffmann 已提交
2077
	    return NULL;
T
ths 已提交
2078
	}
2079
        if (heads < 1 || (type == IF_IDE && heads > 16)) {
G
Gerd Hoffmann 已提交
2080
            fprintf(stderr, "qemu: '%s' invalid physical heads number\n", buf);
G
Gerd Hoffmann 已提交
2081
	    return NULL;
T
ths 已提交
2082
	}
2083
        if (secs < 1 || (type == IF_IDE && secs > 63)) {
G
Gerd Hoffmann 已提交
2084
            fprintf(stderr, "qemu: '%s' invalid physical secs number\n", buf);
G
Gerd Hoffmann 已提交
2085
	    return NULL;
T
ths 已提交
2086 2087 2088
	}
    }

G
Gerd Hoffmann 已提交
2089
    if ((buf = qemu_opt_get(opts, "trans")) != NULL) {
T
ths 已提交
2090 2091 2092
        if (!cyls) {
            fprintf(stderr,
                    "qemu: '%s' trans must be used with cyls,heads and secs\n",
G
Gerd Hoffmann 已提交
2093
                    buf);
G
Gerd Hoffmann 已提交
2094
            return NULL;
T
ths 已提交
2095 2096 2097 2098 2099 2100 2101 2102
        }
        if (!strcmp(buf, "none"))
            translation = BIOS_ATA_TRANSLATION_NONE;
        else if (!strcmp(buf, "lba"))
            translation = BIOS_ATA_TRANSLATION_LBA;
        else if (!strcmp(buf, "auto"))
            translation = BIOS_ATA_TRANSLATION_AUTO;
	else {
G
Gerd Hoffmann 已提交
2103
            fprintf(stderr, "qemu: '%s' invalid translation type\n", buf);
G
Gerd Hoffmann 已提交
2104
	    return NULL;
T
ths 已提交
2105 2106 2107
	}
    }

G
Gerd Hoffmann 已提交
2108
    if ((buf = qemu_opt_get(opts, "media")) != NULL) {
T
ths 已提交
2109 2110 2111 2112 2113
        if (!strcmp(buf, "disk")) {
	    media = MEDIA_DISK;
	} else if (!strcmp(buf, "cdrom")) {
            if (cyls || secs || heads) {
                fprintf(stderr,
G
Gerd Hoffmann 已提交
2114
                        "qemu: '%s' invalid physical CHS format\n", buf);
G
Gerd Hoffmann 已提交
2115
	        return NULL;
T
ths 已提交
2116 2117 2118
            }
	    media = MEDIA_CDROM;
	} else {
G
Gerd Hoffmann 已提交
2119
	    fprintf(stderr, "qemu: '%s' invalid media\n", buf);
G
Gerd Hoffmann 已提交
2120
	    return NULL;
T
ths 已提交
2121 2122 2123
	}
    }

G
Gerd Hoffmann 已提交
2124
    if ((buf = qemu_opt_get(opts, "cache")) != NULL) {
2125
        if (!strcmp(buf, "off") || !strcmp(buf, "none"))
2126
            cache = 0;
2127
        else if (!strcmp(buf, "writethrough"))
2128
            cache = 1;
2129 2130
        else if (!strcmp(buf, "writeback"))
            cache = 2;
2131 2132
        else {
           fprintf(stderr, "qemu: invalid cache option\n");
G
Gerd Hoffmann 已提交
2133
           return NULL;
2134 2135 2136
        }
    }

2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149
#ifdef CONFIG_LINUX_AIO
    if ((buf = qemu_opt_get(opts, "aio")) != NULL) {
        if (!strcmp(buf, "threads"))
            aio = 0;
        else if (!strcmp(buf, "native"))
            aio = 1;
        else {
           fprintf(stderr, "qemu: invalid aio option\n");
           return NULL;
        }
    }
#endif

G
Gerd Hoffmann 已提交
2150
    if ((buf = qemu_opt_get(opts, "format")) != NULL) {
2151 2152 2153 2154
       if (strcmp(buf, "?") == 0) {
            fprintf(stderr, "qemu: Supported formats:");
            bdrv_iterate_format(bdrv_format_print, NULL);
            fprintf(stderr, "\n");
G
Gerd Hoffmann 已提交
2155
	    return NULL;
2156
        }
2157
        drv = bdrv_find_whitelisted_format(buf);
2158 2159
        if (!drv) {
            fprintf(stderr, "qemu: '%s' invalid format\n", buf);
G
Gerd Hoffmann 已提交
2160
            return NULL;
2161 2162 2163
        }
    }

K
Kevin Wolf 已提交
2164
    on_write_error = BLOCK_ERR_STOP_ENOSPC;
G
Gerd Hoffmann 已提交
2165
    if ((buf = qemu_opt_get(opts, "werror")) != NULL) {
2166
        if (type != IF_IDE && type != IF_SCSI && type != IF_VIRTIO) {
2167
            fprintf(stderr, "werror is no supported by this format\n");
G
Gerd Hoffmann 已提交
2168
            return NULL;
2169
        }
K
Kevin Wolf 已提交
2170 2171 2172 2173 2174 2175 2176 2177 2178

        on_write_error = parse_block_error_action(buf, 0);
        if (on_write_error < 0) {
            return NULL;
        }
    }

    on_read_error = BLOCK_ERR_REPORT;
    if ((buf = qemu_opt_get(opts, "rerror")) != NULL) {
K
Kevin Wolf 已提交
2179
        if (type != IF_IDE && type != IF_VIRTIO) {
K
Kevin Wolf 已提交
2180 2181 2182 2183 2184 2185
            fprintf(stderr, "rerror is no supported by this format\n");
            return NULL;
        }

        on_read_error = parse_block_error_action(buf, 1);
        if (on_read_error < 0) {
G
Gerd Hoffmann 已提交
2186
            return NULL;
2187 2188 2189
        }
    }

G
Gerd Hoffmann 已提交
2190
    if ((devaddr = qemu_opt_get(opts, "addr")) != NULL) {
2191
        if (type != IF_VIRTIO) {
G
Gerd Hoffmann 已提交
2192
            fprintf(stderr, "addr is not supported\n");
G
Gerd Hoffmann 已提交
2193
            return NULL;
2194 2195 2196
        }
    }

T
ths 已提交
2197 2198 2199 2200 2201
    /* compute bus and unit according index */

    if (index != -1) {
        if (bus_id != 0 || unit_id != -1) {
            fprintf(stderr,
G
Gerd Hoffmann 已提交
2202
                    "qemu: index cannot be used with bus and unit\n");
G
Gerd Hoffmann 已提交
2203
            return NULL;
T
ths 已提交
2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220
        }
        if (max_devs == 0)
        {
            unit_id = index;
            bus_id = 0;
        } else {
            unit_id = index % max_devs;
            bus_id = index / max_devs;
        }
    }

    /* if user doesn't specify a unit_id,
     * try to find the first free
     */

    if (unit_id == -1) {
       unit_id = 0;
G
Gerd Hoffmann 已提交
2221
       while (drive_get(type, bus_id, unit_id) != NULL) {
T
ths 已提交
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
           unit_id++;
           if (max_devs && unit_id >= max_devs) {
               unit_id -= max_devs;
               bus_id++;
           }
       }
    }

    /* check unit id */

    if (max_devs && unit_id >= max_devs) {
G
Gerd Hoffmann 已提交
2233 2234
        fprintf(stderr, "qemu: unit %d too big (max is %d)\n",
                unit_id, max_devs - 1);
G
Gerd Hoffmann 已提交
2235
        return NULL;
T
ths 已提交
2236 2237 2238 2239 2240 2241
    }

    /*
     * ignore multiple definitions
     */

G
Gerd Hoffmann 已提交
2242 2243 2244 2245
    if (drive_get(type, bus_id, unit_id) != NULL) {
        *fatal_error = 0;
        return NULL;
    }
T
ths 已提交
2246 2247 2248

    /* init */

G
Gerd Hoffmann 已提交
2249
    dinfo = qemu_mallocz(sizeof(*dinfo));
2250
    if ((buf = qemu_opts_id(opts)) != NULL) {
G
Gerd Hoffmann 已提交
2251 2252
        dinfo->id = qemu_strdup(buf);
    } else {
G
Gerd Hoffmann 已提交
2253
        /* no id supplied -> create one */
G
Gerd Hoffmann 已提交
2254
        dinfo->id = qemu_mallocz(32);
G
Gerd Hoffmann 已提交
2255 2256 2257
        if (type == IF_IDE || type == IF_SCSI)
            mediastr = (media == MEDIA_CDROM) ? "-cd" : "-hd";
        if (max_devs)
G
Gerd Hoffmann 已提交
2258
            snprintf(dinfo->id, 32, "%s%i%s%i",
G
Gerd Hoffmann 已提交
2259 2260
                     devname, bus_id, mediastr, unit_id);
        else
G
Gerd Hoffmann 已提交
2261
            snprintf(dinfo->id, 32, "%s%s%i",
G
Gerd Hoffmann 已提交
2262 2263 2264
                     devname, mediastr, unit_id);
    }
    dinfo->bdrv = bdrv_new(dinfo->id);
G
Gerd Hoffmann 已提交
2265 2266 2267 2268
    dinfo->devaddr = devaddr;
    dinfo->type = type;
    dinfo->bus = bus_id;
    dinfo->unit = unit_id;
K
Kevin Wolf 已提交
2269 2270
    dinfo->on_read_error = on_read_error;
    dinfo->on_write_error = on_write_error;
G
Gerd Hoffmann 已提交
2271 2272 2273
    dinfo->opts = opts;
    if (serial)
        strncpy(dinfo->serial, serial, sizeof(serial));
B
Blue Swirl 已提交
2274
    QTAILQ_INSERT_TAIL(&drives, dinfo, next);
T
ths 已提交
2275

2276
    switch(type) {
T
ths 已提交
2277 2278
    case IF_IDE:
    case IF_SCSI:
2279
    case IF_XEN:
G
Gerd Hoffmann 已提交
2280
    case IF_NONE:
T
ths 已提交
2281 2282 2283
        switch(media) {
	case MEDIA_DISK:
            if (cyls != 0) {
G
Gerd Hoffmann 已提交
2284 2285
                bdrv_set_geometry_hint(dinfo->bdrv, cyls, heads, secs);
                bdrv_set_translation_hint(dinfo->bdrv, translation);
T
ths 已提交
2286 2287 2288
            }
	    break;
	case MEDIA_CDROM:
G
Gerd Hoffmann 已提交
2289
            bdrv_set_type_hint(dinfo->bdrv, BDRV_TYPE_CDROM);
T
ths 已提交
2290 2291 2292 2293 2294 2295 2296
	    break;
	}
        break;
    case IF_SD:
        /* FIXME: This isn't really a floppy, but it's a reasonable
           approximation.  */
    case IF_FLOPPY:
G
Gerd Hoffmann 已提交
2297
        bdrv_set_type_hint(dinfo->bdrv, BDRV_TYPE_FLOPPY);
T
ths 已提交
2298 2299 2300 2301
        break;
    case IF_PFLASH:
    case IF_MTD:
        break;
G
Gerd Hoffmann 已提交
2302 2303 2304 2305 2306 2307 2308 2309
    case IF_VIRTIO:
        /* add virtio block device */
        opts = qemu_opts_create(&qemu_device_opts, NULL, 0);
        qemu_opt_set(opts, "driver", "virtio-blk-pci");
        qemu_opt_set(opts, "drive", dinfo->id);
        if (devaddr)
            qemu_opt_set(opts, "addr", devaddr);
        break;
P
Paul Brook 已提交
2310 2311
    case IF_COUNT:
        abort();
T
ths 已提交
2312
    }
G
Gerd Hoffmann 已提交
2313
    if (!file) {
G
Gerd Hoffmann 已提交
2314 2315 2316
        *fatal_error = 0;
        return NULL;
    }
2317
    bdrv_flags = 0;
2318
    if (snapshot) {
2319
        bdrv_flags |= BDRV_O_SNAPSHOT;
2320 2321 2322 2323 2324 2325
        cache = 2; /* always use write-back with snapshot */
    }
    if (cache == 0) /* no caching */
        bdrv_flags |= BDRV_O_NOCACHE;
    else if (cache == 2) /* write-back */
        bdrv_flags |= BDRV_O_CACHE_WB;
2326 2327 2328 2329 2330 2331 2332

    if (aio == 1) {
        bdrv_flags |= BDRV_O_NATIVE_AIO;
    } else {
        bdrv_flags &= ~BDRV_O_NATIVE_AIO;
    }

2333
    if (ro == 1) {
2334 2335
        if (type != IF_SCSI && type != IF_VIRTIO && type != IF_FLOPPY) {
            fprintf(stderr, "qemu: readonly flag not supported for drive with this interface\n");
2336 2337
            return NULL;
        }
N
Naphtali Sprei 已提交
2338 2339 2340 2341 2342 2343
    }
    /* 
     * cdrom is read-only. Set it now, after above interface checking
     * since readonly attribute not explicitly required, so no error.
     */
    if (media == MEDIA_CDROM) {
2344
        ro = 1;
2345
    }
2346
    bdrv_flags |= ro ? 0 : BDRV_O_RDWR;
2347

G
Gerd Hoffmann 已提交
2348
    if (bdrv_open2(dinfo->bdrv, file, bdrv_flags, drv) < 0) {
2349 2350
        fprintf(stderr, "qemu: could not open disk image %s: %s\n",
                        file, strerror(errno));
G
Gerd Hoffmann 已提交
2351
        return NULL;
T
ths 已提交
2352
    }
2353

G
Gerd Hoffmann 已提交
2354
    if (bdrv_key_required(dinfo->bdrv))
2355
        autostart = 0;
G
Gerd Hoffmann 已提交
2356 2357
    *fatal_error = 0;
    return dinfo;
T
ths 已提交
2358 2359
}

G
Gerd Hoffmann 已提交
2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379
static int drive_init_func(QemuOpts *opts, void *opaque)
{
    QEMUMachine *machine = opaque;
    int fatal_error = 0;

    if (drive_init(opts, machine, &fatal_error) == NULL) {
        if (fatal_error)
            return 1;
    }
    return 0;
}

static int drive_enable_snapshot(QemuOpts *opts, void *opaque)
{
    if (NULL == qemu_opt_get(opts, "snapshot")) {
        qemu_opt_set(opts, "snapshot", "on");
    }
    return 0;
}

J
Jan Kiszka 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
void qemu_register_boot_set(QEMUBootSetHandler *func, void *opaque)
{
    boot_set_handler = func;
    boot_set_opaque = opaque;
}

int qemu_boot_set(const char *boot_devices)
{
    if (!boot_set_handler) {
        return -EINVAL;
    }
    return boot_set_handler(boot_set_opaque, boot_devices);
}

J
Jan Kiszka 已提交
2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422
static int parse_bootdevices(char *devices)
{
    /* We just do some generic consistency checks */
    const char *p;
    int bitmap = 0;

    for (p = devices; *p != '\0'; p++) {
        /* Allowed boot devices are:
         * a-b: floppy disk drives
         * c-f: IDE disk drives
         * g-m: machine implementation dependant drives
         * n-p: network devices
         * It's up to each machine implementation to check if the given boot
         * devices match the actual hardware implementation and firmware
         * features.
         */
        if (*p < 'a' || *p > 'p') {
            fprintf(stderr, "Invalid boot device '%c'\n", *p);
            exit(1);
        }
        if (bitmap & (1 << (*p - 'a'))) {
            fprintf(stderr, "Boot device '%c' was given twice\n", *p);
            exit(1);
        }
        bitmap |= 1 << (*p - 'a');
    }
    return bitmap;
}

J
Jan Kiszka 已提交
2423 2424 2425 2426 2427 2428 2429 2430 2431 2432
static void restore_boot_devices(void *opaque)
{
    char *standard_boot_devices = opaque;

    qemu_boot_set(standard_boot_devices);

    qemu_unregister_reset(restore_boot_devices, standard_boot_devices);
    qemu_free(standard_boot_devices);
}

2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476
static void numa_add(const char *optarg)
{
    char option[128];
    char *endptr;
    unsigned long long value, endvalue;
    int nodenr;

    optarg = get_opt_name(option, 128, optarg, ',') + 1;
    if (!strcmp(option, "node")) {
        if (get_param_value(option, 128, "nodeid", optarg) == 0) {
            nodenr = nb_numa_nodes;
        } else {
            nodenr = strtoull(option, NULL, 10);
        }

        if (get_param_value(option, 128, "mem", optarg) == 0) {
            node_mem[nodenr] = 0;
        } else {
            value = strtoull(option, &endptr, 0);
            switch (*endptr) {
            case 0: case 'M': case 'm':
                value <<= 20;
                break;
            case 'G': case 'g':
                value <<= 30;
                break;
            }
            node_mem[nodenr] = value;
        }
        if (get_param_value(option, 128, "cpus", optarg) == 0) {
            node_cpumask[nodenr] = 0;
        } else {
            value = strtoull(option, &endptr, 10);
            if (value >= 64) {
                value = 63;
                fprintf(stderr, "only 64 CPUs in NUMA mode supported.\n");
            } else {
                if (*endptr == '-') {
                    endvalue = strtoull(endptr+1, &endptr, 10);
                    if (endvalue >= 63) {
                        endvalue = 62;
                        fprintf(stderr,
                            "only 63 CPUs in NUMA mode supported.\n");
                    }
P
Paolo Bonzini 已提交
2477
                    value = (2ULL << endvalue) - (1ULL << value);
2478
                } else {
P
Paolo Bonzini 已提交
2479
                    value = 1ULL << value;
2480 2481 2482 2483 2484 2485 2486 2487 2488
                }
            }
            node_cpumask[nodenr] = value;
        }
        nb_numa_nodes++;
    }
    return;
}

2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522
static void smp_parse(const char *optarg)
{
    int smp, sockets = 0, threads = 0, cores = 0;
    char *endptr;
    char option[128];

    smp = strtoul(optarg, &endptr, 10);
    if (endptr != optarg) {
        if (*endptr == ',') {
            endptr++;
        }
    }
    if (get_param_value(option, 128, "sockets", endptr) != 0)
        sockets = strtoull(option, NULL, 10);
    if (get_param_value(option, 128, "cores", endptr) != 0)
        cores = strtoull(option, NULL, 10);
    if (get_param_value(option, 128, "threads", endptr) != 0)
        threads = strtoull(option, NULL, 10);
    if (get_param_value(option, 128, "maxcpus", endptr) != 0)
        max_cpus = strtoull(option, NULL, 10);

    /* compute missing values, prefer sockets over cores over threads */
    if (smp == 0 || sockets == 0) {
        sockets = sockets > 0 ? sockets : 1;
        cores = cores > 0 ? cores : 1;
        threads = threads > 0 ? threads : 1;
        if (smp == 0) {
            smp = cores * threads * sockets;
        }
    } else {
        if (cores == 0) {
            threads = threads > 0 ? threads : 1;
            cores = smp / (sockets * threads);
        } else {
A
Amit Shah 已提交
2523
            if (sockets) {
2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534
                threads = smp / (cores * sockets);
            }
        }
    }
    smp_cpus = smp;
    smp_cores = cores > 0 ? cores : 1;
    smp_threads = threads > 0 ? threads : 1;
    if (max_cpus == 0)
        max_cpus = smp_cpus;
}

B
bellard 已提交
2535 2536 2537
/***********************************************************/
/* USB devices */

2538
static int usb_device_add(const char *devname, int is_hotplug)
B
bellard 已提交
2539 2540
{
    const char *p;
2541
    USBDevice *dev = NULL;
B
bellard 已提交
2542

2543
    if (!usb_enabled)
B
bellard 已提交
2544 2545
        return -1;

G
Gerd Hoffmann 已提交
2546 2547 2548 2549 2550
    /* drivers with .usbdevice_name entry in USBDeviceInfo */
    dev = usbdevice_create(devname);
    if (dev)
        goto done;

2551
    /* the other ones */
B
bellard 已提交
2552 2553
    if (strstart(devname, "host:", &p)) {
        dev = usb_host_device_open(p);
2554 2555 2556
    } else if (!strcmp(devname, "bt") || strstart(devname, "bt:", &p)) {
        dev = usb_bt_init(devname[2] ? hci_init(p) :
                        bt_new_hci(qemu_find_bt_vlan(0)));
B
bellard 已提交
2557 2558 2559
    } else {
        return -1;
    }
P
pbrook 已提交
2560 2561 2562
    if (!dev)
        return -1;

2563
done:
B
bellard 已提交
2564 2565 2566
    return 0;
}

2567 2568 2569 2570 2571
static int usb_device_del(const char *devname)
{
    int bus_num, addr;
    const char *p;

2572 2573 2574
    if (strstart(devname, "host:", &p))
        return usb_host_device_close(p);

2575
    if (!usb_enabled)
2576 2577 2578 2579 2580 2581 2582 2583
        return -1;

    p = strchr(devname, '.');
    if (!p)
        return -1;
    bus_num = strtoul(devname, NULL, 0);
    addr = strtoul(p + 1, NULL, 0);

2584
    return usb_device_delete_addr(bus_num, addr);
2585 2586
}

2587 2588
static int usb_parse(const char *cmdline)
{
S
Scott Tsai 已提交
2589 2590 2591 2592 2593 2594
    int r;
    r = usb_device_add(cmdline, 0);
    if (r < 0) {
        fprintf(stderr, "qemu: could not add USB device '%s'\n", cmdline);
    }
    return r;
2595 2596
}

2597
void do_usb_add(Monitor *mon, const QDict *qdict)
B
bellard 已提交
2598
{
S
Scott Tsai 已提交
2599 2600
    const char *devname = qdict_get_str(qdict, "devname");
    if (usb_device_add(devname, 1) < 0) {
2601
        error_report("could not add USB device '%s'", devname);
S
Scott Tsai 已提交
2602
    }
B
bellard 已提交
2603 2604
}

2605
void do_usb_del(Monitor *mon, const QDict *qdict)
B
bellard 已提交
2606
{
S
Scott Tsai 已提交
2607 2608
    const char *devname = qdict_get_str(qdict, "devname");
    if (usb_device_del(devname) < 0) {
2609
        error_report("could not delete USB device '%s'", devname);
S
Scott Tsai 已提交
2610
    }
B
bellard 已提交
2611 2612
}

2613 2614 2615 2616
/***********************************************************/
/* PCMCIA/Cardbus */

static struct pcmcia_socket_entry_s {
P
Paul Brook 已提交
2617
    PCMCIASocket *socket;
2618 2619 2620
    struct pcmcia_socket_entry_s *next;
} *pcmcia_sockets = 0;

P
Paul Brook 已提交
2621
void pcmcia_socket_register(PCMCIASocket *socket)
2622 2623 2624 2625 2626 2627 2628 2629 2630
{
    struct pcmcia_socket_entry_s *entry;

    entry = qemu_malloc(sizeof(struct pcmcia_socket_entry_s));
    entry->socket = socket;
    entry->next = pcmcia_sockets;
    pcmcia_sockets = entry;
}

P
Paul Brook 已提交
2631
void pcmcia_socket_unregister(PCMCIASocket *socket)
2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642
{
    struct pcmcia_socket_entry_s *entry, **ptr;

    ptr = &pcmcia_sockets;
    for (entry = *ptr; entry; ptr = &entry->next, entry = *ptr)
        if (entry->socket == socket) {
            *ptr = entry->next;
            qemu_free(entry);
        }
}

A
aliguori 已提交
2643
void pcmcia_info(Monitor *mon)
2644 2645
{
    struct pcmcia_socket_entry_s *iter;
A
aliguori 已提交
2646

2647
    if (!pcmcia_sockets)
A
aliguori 已提交
2648
        monitor_printf(mon, "No PCMCIA sockets\n");
2649 2650

    for (iter = pcmcia_sockets; iter; iter = iter->next)
A
aliguori 已提交
2651 2652 2653
        monitor_printf(mon, "%s: %s\n", iter->socket->slot_string,
                       iter->socket->attached ? iter->socket->card_string :
                       "Empty");
2654 2655
}

2656
/***********************************************************/
2657
/* I/O handling */
2658

2659 2660
typedef struct IOHandlerRecord {
    int fd;
B
bellard 已提交
2661 2662 2663
    IOCanRWHandler *fd_read_poll;
    IOHandler *fd_read;
    IOHandler *fd_write;
2664
    int deleted;
2665 2666 2667
    void *opaque;
    /* temporary data */
    struct pollfd *ufd;
2668
    struct IOHandlerRecord *next;
2669 2670
} IOHandlerRecord;

2671
static IOHandlerRecord *first_io_handler;
2672

B
bellard 已提交
2673 2674
/* XXX: fd_read_poll should be suppressed, but an API change is
   necessary in the character devices to suppress fd_can_read(). */
2675 2676 2677 2678
int qemu_set_fd_handler2(int fd,
                         IOCanRWHandler *fd_read_poll,
                         IOHandler *fd_read,
                         IOHandler *fd_write,
B
bellard 已提交
2679
                         void *opaque)
2680
{
B
bellard 已提交
2681
    IOHandlerRecord **pioh, *ioh;
2682

B
bellard 已提交
2683 2684 2685 2686 2687 2688 2689
    if (!fd_read && !fd_write) {
        pioh = &first_io_handler;
        for(;;) {
            ioh = *pioh;
            if (ioh == NULL)
                break;
            if (ioh->fd == fd) {
2690
                ioh->deleted = 1;
B
bellard 已提交
2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708
                break;
            }
            pioh = &ioh->next;
        }
    } else {
        for(ioh = first_io_handler; ioh != NULL; ioh = ioh->next) {
            if (ioh->fd == fd)
                goto found;
        }
        ioh = qemu_mallocz(sizeof(IOHandlerRecord));
        ioh->next = first_io_handler;
        first_io_handler = ioh;
    found:
        ioh->fd = fd;
        ioh->fd_read_poll = fd_read_poll;
        ioh->fd_read = fd_read;
        ioh->fd_write = fd_write;
        ioh->opaque = opaque;
2709
        ioh->deleted = 0;
B
bellard 已提交
2710
    }
2711 2712 2713
    return 0;
}

2714 2715 2716
int qemu_set_fd_handler(int fd,
                        IOHandler *fd_read,
                        IOHandler *fd_write,
B
bellard 已提交
2717
                        void *opaque)
2718
{
B
bellard 已提交
2719
    return qemu_set_fd_handler2(fd, NULL, fd_read, fd_write, opaque);
2720 2721
}

A
aliguori 已提交
2722
#ifdef _WIN32
2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757
/***********************************************************/
/* Polling handling */

typedef struct PollingEntry {
    PollingFunc *func;
    void *opaque;
    struct PollingEntry *next;
} PollingEntry;

static PollingEntry *first_polling_entry;

int qemu_add_polling_cb(PollingFunc *func, void *opaque)
{
    PollingEntry **ppe, *pe;
    pe = qemu_mallocz(sizeof(PollingEntry));
    pe->func = func;
    pe->opaque = opaque;
    for(ppe = &first_polling_entry; *ppe != NULL; ppe = &(*ppe)->next);
    *ppe = pe;
    return 0;
}

void qemu_del_polling_cb(PollingFunc *func, void *opaque)
{
    PollingEntry **ppe, *pe;
    for(ppe = &first_polling_entry; *ppe != NULL; ppe = &(*ppe)->next) {
        pe = *ppe;
        if (pe->func == func && pe->opaque == opaque) {
            *ppe = pe->next;
            qemu_free(pe);
            break;
        }
    }
}

2758 2759 2760 2761 2762 2763 2764 2765 2766 2767
/***********************************************************/
/* Wait objects support */
typedef struct WaitObjects {
    int num;
    HANDLE events[MAXIMUM_WAIT_OBJECTS + 1];
    WaitObjectFunc *func[MAXIMUM_WAIT_OBJECTS + 1];
    void *opaque[MAXIMUM_WAIT_OBJECTS + 1];
} WaitObjects;

static WaitObjects wait_objects = {0};
2768

2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794
int qemu_add_wait_object(HANDLE handle, WaitObjectFunc *func, void *opaque)
{
    WaitObjects *w = &wait_objects;

    if (w->num >= MAXIMUM_WAIT_OBJECTS)
        return -1;
    w->events[w->num] = handle;
    w->func[w->num] = func;
    w->opaque[w->num] = opaque;
    w->num++;
    return 0;
}

void qemu_del_wait_object(HANDLE handle, WaitObjectFunc *func, void *opaque)
{
    int i, found;
    WaitObjects *w = &wait_objects;

    found = 0;
    for (i = 0; i < w->num; i++) {
        if (w->events[i] == handle)
            found = 1;
        if (found) {
            w->events[i] = w->events[i + 1];
            w->func[i] = w->func[i + 1];
            w->opaque[i] = w->opaque[i + 1];
2795
        }
2796 2797 2798 2799 2800 2801
    }
    if (found)
        w->num--;
}
#endif

2802 2803 2804
/***********************************************************/
/* ram save/restore */

2805
#define RAM_SAVE_FLAG_FULL	0x01 /* Obsolete, not used anymore */
2806 2807 2808 2809 2810 2811
#define RAM_SAVE_FLAG_COMPRESS	0x02
#define RAM_SAVE_FLAG_MEM_SIZE	0x04
#define RAM_SAVE_FLAG_PAGE	0x08
#define RAM_SAVE_FLAG_EOS	0x10

static int is_dup_page(uint8_t *page, uint8_t ch)
2812
{
2813 2814 2815
    uint32_t val = ch << 24 | ch << 16 | ch << 8 | ch;
    uint32_t *array = (uint32_t *)page;
    int i;
2816

2817 2818 2819 2820 2821 2822 2823 2824 2825 2826
    for (i = 0; i < (TARGET_PAGE_SIZE / 4); i++) {
        if (array[i] != val)
            return 0;
    }

    return 1;
}

static int ram_save_block(QEMUFile *f)
{
A
Anthony Liguori 已提交
2827 2828 2829
    static ram_addr_t current_addr = 0;
    ram_addr_t saved_addr = current_addr;
    ram_addr_t addr = 0;
2830 2831
    int found = 0;

P
pbrook 已提交
2832
    while (addr < last_ram_offset) {
2833
        if (cpu_physical_memory_get_dirty(current_addr, MIGRATION_DIRTY_FLAG)) {
P
pbrook 已提交
2834
            uint8_t *p;
2835 2836 2837 2838 2839

            cpu_physical_memory_reset_dirty(current_addr,
                                            current_addr + TARGET_PAGE_SIZE,
                                            MIGRATION_DIRTY_FLAG);

P
pbrook 已提交
2840
            p = qemu_get_ram_ptr(current_addr);
2841

P
pbrook 已提交
2842
            if (is_dup_page(p, *p)) {
2843
                qemu_put_be64(f, current_addr | RAM_SAVE_FLAG_COMPRESS);
P
pbrook 已提交
2844
                qemu_put_byte(f, *p);
2845 2846
            } else {
                qemu_put_be64(f, current_addr | RAM_SAVE_FLAG_PAGE);
P
pbrook 已提交
2847
                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
2848
            }
2849 2850 2851

            found = 1;
            break;
2852
        }
2853
        addr += TARGET_PAGE_SIZE;
P
pbrook 已提交
2854
        current_addr = (saved_addr + addr) % last_ram_offset;
2855
    }
2856 2857

    return found;
2858 2859
}

2860
static uint64_t bytes_transferred;
2861

A
Anthony Liguori 已提交
2862
static ram_addr_t ram_save_remaining(void)
2863
{
A
Anthony Liguori 已提交
2864 2865
    ram_addr_t addr;
    ram_addr_t count = 0;
2866

P
pbrook 已提交
2867
    for (addr = 0; addr < last_ram_offset; addr += TARGET_PAGE_SIZE) {
2868 2869 2870 2871 2872 2873 2874
        if (cpu_physical_memory_get_dirty(addr, MIGRATION_DIRTY_FLAG))
            count++;
    }

    return count;
}

2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889
uint64_t ram_bytes_remaining(void)
{
    return ram_save_remaining() * TARGET_PAGE_SIZE;
}

uint64_t ram_bytes_transferred(void)
{
    return bytes_transferred;
}

uint64_t ram_bytes_total(void)
{
    return last_ram_offset;
}

2890
static int ram_save_live(Monitor *mon, QEMUFile *f, int stage, void *opaque)
2891
{
A
Anthony Liguori 已提交
2892
    ram_addr_t addr;
2893 2894 2895
    uint64_t bytes_transferred_last;
    double bwidth = 0;
    uint64_t expected_time = 0;
2896

2897 2898 2899 2900 2901
    if (stage < 0) {
        cpu_physical_memory_set_dirty_tracking(0);
        return 0;
    }

2902
    if (cpu_physical_sync_dirty_bitmap(0, TARGET_PHYS_ADDR_MAX) != 0) {
2903 2904 2905 2906
        qemu_file_set_error(f);
        return 0;
    }

2907
    if (stage == 1) {
2908 2909
        bytes_transferred = 0;

2910
        /* Make sure all dirty bits are set */
P
pbrook 已提交
2911
        for (addr = 0; addr < last_ram_offset; addr += TARGET_PAGE_SIZE) {
2912 2913 2914
            if (!cpu_physical_memory_get_dirty(addr, MIGRATION_DIRTY_FLAG))
                cpu_physical_memory_set_dirty(addr);
        }
2915

2916 2917 2918
        /* Enable dirty memory tracking */
        cpu_physical_memory_set_dirty_tracking(1);

P
pbrook 已提交
2919
        qemu_put_be64(f, last_ram_offset | RAM_SAVE_FLAG_MEM_SIZE);
2920 2921
    }

2922
    bytes_transferred_last = bytes_transferred;
P
Paolo Bonzini 已提交
2923
    bwidth = qemu_get_clock_ns(rt_clock);
2924

2925 2926 2927 2928
    while (!qemu_file_rate_limit(f)) {
        int ret;

        ret = ram_save_block(f);
2929
        bytes_transferred += ret * TARGET_PAGE_SIZE;
2930 2931 2932 2933
        if (ret == 0) /* no more blocks */
            break;
    }

P
Paolo Bonzini 已提交
2934
    bwidth = qemu_get_clock_ns(rt_clock) - bwidth;
2935 2936 2937 2938 2939 2940 2941
    bwidth = (bytes_transferred - bytes_transferred_last) / bwidth;

    /* if we haven't transferred anything this round, force expected_time to a
     * a very high value, but without crashing */
    if (bwidth == 0)
        bwidth = 0.000001;

2942 2943 2944
    /* try transferring iterative blocks of memory */
    if (stage == 3) {
        /* flush all remaining blocks regardless of rate limiting */
2945 2946 2947
        while (ram_save_block(f) != 0) {
            bytes_transferred += TARGET_PAGE_SIZE;
        }
2948
        cpu_physical_memory_set_dirty_tracking(0);
2949 2950 2951 2952
    }

    qemu_put_be64(f, RAM_SAVE_FLAG_EOS);

2953 2954 2955
    expected_time = ram_save_remaining() * TARGET_PAGE_SIZE / bwidth;

    return (stage == 2) && (expected_time <= migrate_max_downtime());
2956 2957 2958 2959
}

static int ram_load(QEMUFile *f, void *opaque, int version_id)
{
A
Anthony Liguori 已提交
2960
    ram_addr_t addr;
2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972
    int flags;

    if (version_id != 3)
        return -EINVAL;

    do {
        addr = qemu_get_be64(f);

        flags = addr & ~TARGET_PAGE_MASK;
        addr &= TARGET_PAGE_MASK;

        if (flags & RAM_SAVE_FLAG_MEM_SIZE) {
P
pbrook 已提交
2973
            if (addr != last_ram_offset)
2974 2975 2976 2977 2978
                return -EINVAL;
        }

        if (flags & RAM_SAVE_FLAG_COMPRESS) {
            uint8_t ch = qemu_get_byte(f);
2979 2980
            memset(qemu_get_ram_ptr(addr), ch, TARGET_PAGE_SIZE);
#ifndef _WIN32
2981 2982 2983
            if (ch == 0 &&
                (!kvm_enabled() || kvm_has_sync_mmu())) {
                madvise(qemu_get_ram_ptr(addr), TARGET_PAGE_SIZE, MADV_DONTNEED);
2984
            }
2985
#endif
J
Jan Kiszka 已提交
2986
        } else if (flags & RAM_SAVE_FLAG_PAGE) {
P
pbrook 已提交
2987
            qemu_get_buffer(f, qemu_get_ram_ptr(addr), TARGET_PAGE_SIZE);
J
Jan Kiszka 已提交
2988 2989 2990 2991
        }
        if (qemu_file_has_error(f)) {
            return -EIO;
        }
2992 2993
    } while (!(flags & RAM_SAVE_FLAG_EOS));

2994 2995 2996
    return 0;
}

A
aliguori 已提交
2997 2998
void qemu_service_io(void)
{
2999
    qemu_notify_event();
A
aliguori 已提交
3000 3001
}

3002 3003 3004
/***********************************************************/
/* machine registration */

B
blueswir1 已提交
3005
static QEMUMachine *first_machine = NULL;
3006
QEMUMachine *current_machine = NULL;
3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018

int qemu_register_machine(QEMUMachine *m)
{
    QEMUMachine **pm;
    pm = &first_machine;
    while (*pm != NULL)
        pm = &(*pm)->next;
    m->next = NULL;
    *pm = m;
    return 0;
}

3019
static QEMUMachine *find_machine(const char *name)
3020 3021 3022 3023 3024 3025
{
    QEMUMachine *m;

    for(m = first_machine; m != NULL; m = m->next) {
        if (!strcmp(m->name, name))
            return m;
M
Mark McLoughlin 已提交
3026 3027
        if (m->alias && !strcmp(m->alias, name))
            return m;
3028 3029 3030 3031
    }
    return NULL;
}

3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043
static QEMUMachine *find_default_machine(void)
{
    QEMUMachine *m;

    for(m = first_machine; m != NULL; m = m->next) {
        if (m->is_default) {
            return m;
        }
    }
    return NULL;
}

3044 3045 3046
/***********************************************************/
/* main execution loop */

3047
static void gui_update(void *opaque)
3048
{
3049
    uint64_t interval = GUI_REFRESH_INTERVAL;
3050
    DisplayState *ds = opaque;
3051 3052
    DisplayChangeListener *dcl = ds->listeners;

3053
    qemu_flush_coalesced_mmio_buffer();
3054 3055 3056 3057 3058 3059 3060 3061 3062
    dpy_refresh(ds);

    while (dcl != NULL) {
        if (dcl->gui_timer_interval &&
            dcl->gui_timer_interval < interval)
            interval = dcl->gui_timer_interval;
        dcl = dcl->next;
    }
    qemu_mod_timer(ds->gui_timer, interval + qemu_get_clock(rt_clock));
3063 3064
}

B
blueswir1 已提交
3065 3066 3067 3068
static void nographic_update(void *opaque)
{
    uint64_t interval = GUI_REFRESH_INTERVAL;

3069
    qemu_flush_coalesced_mmio_buffer();
B
blueswir1 已提交
3070 3071 3072
    qemu_mod_timer(nographic_timer, interval + qemu_get_clock(rt_clock));
}

3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099
void cpu_synchronize_all_states(void)
{
    CPUState *cpu;

    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
        cpu_synchronize_state(cpu);
    }
}

void cpu_synchronize_all_post_reset(void)
{
    CPUState *cpu;

    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
        cpu_synchronize_post_reset(cpu);
    }
}

void cpu_synchronize_all_post_init(void)
{
    CPUState *cpu;

    for (cpu = first_cpu; cpu; cpu = cpu->next_cpu) {
        cpu_synchronize_post_init(cpu);
    }
}

B
bellard 已提交
3100 3101 3102
struct vm_change_state_entry {
    VMChangeStateHandler *cb;
    void *opaque;
B
Blue Swirl 已提交
3103
    QLIST_ENTRY (vm_change_state_entry) entries;
B
bellard 已提交
3104 3105
};

B
Blue Swirl 已提交
3106
static QLIST_HEAD(vm_change_state_head, vm_change_state_entry) vm_change_state_head;
B
bellard 已提交
3107 3108 3109 3110 3111 3112 3113 3114 3115 3116

VMChangeStateEntry *qemu_add_vm_change_state_handler(VMChangeStateHandler *cb,
                                                     void *opaque)
{
    VMChangeStateEntry *e;

    e = qemu_mallocz(sizeof (*e));

    e->cb = cb;
    e->opaque = opaque;
B
Blue Swirl 已提交
3117
    QLIST_INSERT_HEAD(&vm_change_state_head, e, entries);
B
bellard 已提交
3118 3119 3120 3121 3122
    return e;
}

void qemu_del_vm_change_state_handler(VMChangeStateEntry *e)
{
B
Blue Swirl 已提交
3123
    QLIST_REMOVE (e, entries);
B
bellard 已提交
3124 3125 3126
    qemu_free (e);
}

3127
static void vm_state_notify(int running, int reason)
B
bellard 已提交
3128 3129 3130 3131
{
    VMChangeStateEntry *e;

    for (e = vm_change_state_head.lh_first; e; e = e->entries.le_next) {
3132
        e->cb(e->opaque, running, reason);
B
bellard 已提交
3133 3134 3135
    }
}

3136 3137 3138
static void resume_all_vcpus(void);
static void pause_all_vcpus(void);

3139 3140 3141 3142 3143
void vm_start(void)
{
    if (!vm_running) {
        cpu_enable_ticks();
        vm_running = 1;
3144
        vm_state_notify(1, 0);
3145
        resume_all_vcpus();
3146 3147 3148
    }
}

3149 3150 3151
/* reset/shutdown handler */

typedef struct QEMUResetEntry {
B
Blue Swirl 已提交
3152
    QTAILQ_ENTRY(QEMUResetEntry) entry;
3153 3154 3155 3156
    QEMUResetHandler *func;
    void *opaque;
} QEMUResetEntry;

B
Blue Swirl 已提交
3157 3158
static QTAILQ_HEAD(reset_handlers, QEMUResetEntry) reset_handlers =
    QTAILQ_HEAD_INITIALIZER(reset_handlers);
3159 3160
static int reset_requested;
static int shutdown_requested;
B
bellard 已提交
3161
static int powerdown_requested;
3162
static int debug_requested;
3163
static int vmstop_requested;
3164

A
aurel32 已提交
3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185
int qemu_shutdown_requested(void)
{
    int r = shutdown_requested;
    shutdown_requested = 0;
    return r;
}

int qemu_reset_requested(void)
{
    int r = reset_requested;
    reset_requested = 0;
    return r;
}

int qemu_powerdown_requested(void)
{
    int r = powerdown_requested;
    powerdown_requested = 0;
    return r;
}

3186 3187 3188 3189 3190 3191 3192
static int qemu_debug_requested(void)
{
    int r = debug_requested;
    debug_requested = 0;
    return r;
}

3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204
static int qemu_vmstop_requested(void)
{
    int r = vmstop_requested;
    vmstop_requested = 0;
    return r;
}

static void do_vm_stop(int reason)
{
    if (vm_running) {
        cpu_disable_ticks();
        vm_running = 0;
3205
        pause_all_vcpus();
3206
        vm_state_notify(0, reason);
3207
        monitor_protocol_event(QEVENT_STOP, NULL);
3208 3209 3210
    }
}

3211
void qemu_register_reset(QEMUResetHandler *func, void *opaque)
3212
{
J
Jan Kiszka 已提交
3213
    QEMUResetEntry *re = qemu_mallocz(sizeof(QEMUResetEntry));
3214 3215 3216

    re->func = func;
    re->opaque = opaque;
B
Blue Swirl 已提交
3217
    QTAILQ_INSERT_TAIL(&reset_handlers, re, entry);
3218 3219
}

J
Jan Kiszka 已提交
3220
void qemu_unregister_reset(QEMUResetHandler *func, void *opaque)
3221 3222 3223
{
    QEMUResetEntry *re;

B
Blue Swirl 已提交
3224
    QTAILQ_FOREACH(re, &reset_handlers, entry) {
J
Jan Kiszka 已提交
3225
        if (re->func == func && re->opaque == opaque) {
B
Blue Swirl 已提交
3226
            QTAILQ_REMOVE(&reset_handlers, re, entry);
J
Jan Kiszka 已提交
3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237
            qemu_free(re);
            return;
        }
    }
}

void qemu_system_reset(void)
{
    QEMUResetEntry *re, *nre;

    /* reset all devices */
B
Blue Swirl 已提交
3238
    QTAILQ_FOREACH_SAFE(re, &reset_handlers, entry, nre) {
3239 3240
        re->func(re->opaque);
    }
3241
    monitor_protocol_event(QEVENT_RESET, NULL);
3242
    cpu_synchronize_all_post_reset();
3243 3244 3245 3246
}

void qemu_system_reset_request(void)
{
B
bellard 已提交
3247 3248 3249 3250 3251
    if (no_reboot) {
        shutdown_requested = 1;
    } else {
        reset_requested = 1;
    }
3252
    qemu_notify_event();
3253 3254 3255 3256 3257
}

void qemu_system_shutdown_request(void)
{
    shutdown_requested = 1;
3258
    qemu_notify_event();
3259 3260
}

B
bellard 已提交
3261 3262 3263
void qemu_system_powerdown_request(void)
{
    powerdown_requested = 1;
3264 3265 3266
    qemu_notify_event();
}

3267 3268
#ifdef CONFIG_IOTHREAD
static void qemu_system_vmstop_request(int reason)
3269
{
3270 3271
    vmstop_requested = reason;
    qemu_notify_event();
3272
}
3273
#endif
3274

3275 3276 3277 3278
#ifndef _WIN32
static int io_thread_fd = -1;

static void qemu_event_increment(void)
3279
{
P
Paolo Bonzini 已提交
3280 3281
    /* Write 8 bytes to be compatible with eventfd.  */
    static uint64_t val = 1;
3282
    ssize_t ret;
3283 3284 3285 3286

    if (io_thread_fd == -1)
        return;

3287
    do {
P
Paolo Bonzini 已提交
3288
        ret = write(io_thread_fd, &val, sizeof(val));
3289 3290 3291 3292
    } while (ret < 0 && errno == EINTR);

    /* EAGAIN is fine, a read must be pending.  */
    if (ret < 0 && errno != EAGAIN) {
3293 3294 3295 3296
        fprintf(stderr, "qemu_event_increment: write() filed: %s\n",
                strerror(errno));
        exit (1);
    }
3297 3298 3299 3300 3301 3302
}

static void qemu_event_read(void *opaque)
{
    int fd = (unsigned long)opaque;
    ssize_t len;
3303
    char buffer[512];
3304

P
Paolo Bonzini 已提交
3305
    /* Drain the notify pipe.  For eventfd, only 8 bytes will be read.  */
3306 3307
    do {
        len = read(fd, buffer, sizeof(buffer));
3308
    } while ((len == -1 && errno == EINTR) || len == sizeof(buffer));
3309 3310 3311 3312 3313 3314 3315
}

static int qemu_event_init(void)
{
    int err;
    int fds[2];

P
Paolo Bonzini 已提交
3316
    err = qemu_eventfd(fds);
3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331
    if (err == -1)
        return -errno;

    err = fcntl_setfl(fds[0], O_NONBLOCK);
    if (err < 0)
        goto fail;

    err = fcntl_setfl(fds[1], O_NONBLOCK);
    if (err < 0)
        goto fail;

    qemu_set_fd_handler2(fds[0], NULL, qemu_event_read, NULL,
                         (void *)(unsigned long)fds[0]);

    io_thread_fd = fds[1];
J
Jan Kiszka 已提交
3332 3333
    return 0;

3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349
fail:
    close(fds[0]);
    close(fds[1]);
    return err;
}
#else
HANDLE qemu_event_handle;

static void dummy_event_handler(void *opaque)
{
}

static int qemu_event_init(void)
{
    qemu_event_handle = CreateEvent(NULL, FALSE, FALSE, NULL);
    if (!qemu_event_handle) {
B
Blue Swirl 已提交
3350
        fprintf(stderr, "Failed CreateEvent: %ld\n", GetLastError());
3351 3352 3353
        return -1;
    }
    qemu_add_wait_object(qemu_event_handle, dummy_event_handler, NULL);
3354 3355 3356
    return 0;
}

3357 3358
static void qemu_event_increment(void)
{
M
malc 已提交
3359
    if (!SetEvent(qemu_event_handle)) {
B
Blue Swirl 已提交
3360
        fprintf(stderr, "qemu_event_increment: SetEvent failed: %ld\n",
M
malc 已提交
3361 3362 3363
                GetLastError());
        exit (1);
    }
3364 3365 3366
}
#endif

3367 3368 3369 3370 3371 3372
static int cpu_can_run(CPUState *env)
{
    if (env->stop)
        return 0;
    if (env->stopped)
        return 0;
3373 3374
    if (!vm_running)
        return 0;
3375 3376 3377 3378
    return 1;
}

#ifndef CONFIG_IOTHREAD
3379 3380 3381 3382 3383
static int qemu_init_main_loop(void)
{
    return qemu_event_init();
}

3384 3385 3386 3387
void qemu_init_vcpu(void *_env)
{
    CPUState *env = _env;

3388 3389
    env->nr_cores = smp_cores;
    env->nr_threads = smp_threads;
J
Jiri Denemark 已提交
3390 3391
    if (kvm_enabled())
        kvm_init_vcpu(env);
3392 3393 3394
    return;
}

3395 3396 3397 3398 3399
int qemu_cpu_self(void *env)
{
    return 1;
}

3400 3401 3402 3403 3404 3405 3406 3407
static void resume_all_vcpus(void)
{
}

static void pause_all_vcpus(void)
{
}

3408 3409 3410 3411 3412
void qemu_cpu_kick(void *env)
{
    return;
}

3413 3414 3415 3416
void qemu_notify_event(void)
{
    CPUState *env = cpu_single_env;

3417
    qemu_event_increment ();
3418 3419
    if (env) {
        cpu_exit(env);
3420
    }
P
Paolo Bonzini 已提交
3421 3422 3423
    if (next_cpu && env != next_cpu) {
	cpu_exit(next_cpu);
    }
3424 3425
}

3426 3427
void qemu_mutex_lock_iothread(void) {}
void qemu_mutex_unlock_iothread(void) {}
3428

3429 3430 3431 3432 3433
void vm_stop(int reason)
{
    do_vm_stop(reason);
}

3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452
#else /* CONFIG_IOTHREAD */

#include "qemu-thread.h"

QemuMutex qemu_global_mutex;
static QemuMutex qemu_fair_mutex;

static QemuThread io_thread;

static QemuThread *tcg_cpu_thread;
static QemuCond *tcg_halt_cond;

static int qemu_system_ready;
/* cpu creation */
static QemuCond qemu_cpu_cond;
/* system init */
static QemuCond qemu_system_cond;
static QemuCond qemu_pause_cond;

3453 3454
static void tcg_block_io_signals(void);
static void kvm_block_io_signals(CPUState *env);
3455 3456
static void unblock_io_signals(void);
static int tcg_has_work(void);
M
Marcelo Tosatti 已提交
3457
static int cpu_has_work(CPUState *env);
3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477

static int qemu_init_main_loop(void)
{
    int ret;

    ret = qemu_event_init();
    if (ret)
        return ret;

    qemu_cond_init(&qemu_pause_cond);
    qemu_mutex_init(&qemu_fair_mutex);
    qemu_mutex_init(&qemu_global_mutex);
    qemu_mutex_lock(&qemu_global_mutex);

    unblock_io_signals();
    qemu_thread_self(&io_thread);

    return 0;
}

M
Marcelo Tosatti 已提交
3478 3479 3480 3481 3482 3483 3484 3485 3486
static void qemu_wait_io_event_common(CPUState *env)
{
    if (env->stop) {
        env->stop = 0;
        env->stopped = 1;
        qemu_cond_signal(&qemu_pause_cond);
    }
}

3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502
static void qemu_wait_io_event(CPUState *env)
{
    while (!tcg_has_work())
        qemu_cond_timedwait(env->halt_cond, &qemu_global_mutex, 1000);

    qemu_mutex_unlock(&qemu_global_mutex);

    /*
     * Users of qemu_global_mutex can be starved, having no chance
     * to acquire it since this path will get to it first.
     * So use another lock to provide fairness.
     */
    qemu_mutex_lock(&qemu_fair_mutex);
    qemu_mutex_unlock(&qemu_fair_mutex);

    qemu_mutex_lock(&qemu_global_mutex);
M
Marcelo Tosatti 已提交
3503 3504 3505
    qemu_wait_io_event_common(env);
}

3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526
static void qemu_kvm_eat_signal(CPUState *env, int timeout)
{
    struct timespec ts;
    int r, e;
    siginfo_t siginfo;
    sigset_t waitset;

    ts.tv_sec = timeout / 1000;
    ts.tv_nsec = (timeout % 1000) * 1000000;

    sigemptyset(&waitset);
    sigaddset(&waitset, SIG_IPI);

    qemu_mutex_unlock(&qemu_global_mutex);
    r = sigtimedwait(&waitset, &siginfo, &ts);
    e = errno;
    qemu_mutex_lock(&qemu_global_mutex);

    if (r == -1 && !(e == EAGAIN || e == EINTR)) {
        fprintf(stderr, "sigtimedwait: %s\n", strerror(e));
        exit(1);
3527 3528 3529
    }
}

M
Marcelo Tosatti 已提交
3530 3531 3532 3533 3534
static void qemu_kvm_wait_io_event(CPUState *env)
{
    while (!cpu_has_work(env))
        qemu_cond_timedwait(env->halt_cond, &qemu_global_mutex, 1000);

3535
    qemu_kvm_eat_signal(env, 0);
M
Marcelo Tosatti 已提交
3536
    qemu_wait_io_event_common(env);
3537 3538 3539 3540 3541 3542 3543 3544 3545
}

static int qemu_cpu_exec(CPUState *env);

static void *kvm_cpu_thread_fn(void *arg)
{
    CPUState *env = arg;

    qemu_thread_self(env->thread);
3546 3547
    if (kvm_enabled())
        kvm_init_vcpu(env);
3548

3549 3550
    kvm_block_io_signals(env);

3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562
    /* signal CPU creation */
    qemu_mutex_lock(&qemu_global_mutex);
    env->created = 1;
    qemu_cond_signal(&qemu_cpu_cond);

    /* and wait for machine initialization */
    while (!qemu_system_ready)
        qemu_cond_timedwait(&qemu_system_cond, &qemu_global_mutex, 100);

    while (1) {
        if (cpu_can_run(env))
            qemu_cpu_exec(env);
M
Marcelo Tosatti 已提交
3563
        qemu_kvm_wait_io_event(env);
3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574
    }

    return NULL;
}

static void tcg_cpu_exec(void);

static void *tcg_cpu_thread_fn(void *arg)
{
    CPUState *env = arg;

3575
    tcg_block_io_signals();
3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600
    qemu_thread_self(env->thread);

    /* signal CPU creation */
    qemu_mutex_lock(&qemu_global_mutex);
    for (env = first_cpu; env != NULL; env = env->next_cpu)
        env->created = 1;
    qemu_cond_signal(&qemu_cpu_cond);

    /* and wait for machine initialization */
    while (!qemu_system_ready)
        qemu_cond_timedwait(&qemu_system_cond, &qemu_global_mutex, 100);

    while (1) {
        tcg_cpu_exec();
        qemu_wait_io_event(cur_cpu);
    }

    return NULL;
}

void qemu_cpu_kick(void *_env)
{
    CPUState *env = _env;
    qemu_cond_broadcast(env->halt_cond);
    if (kvm_enabled())
3601
        qemu_thread_signal(env->thread, SIG_IPI);
3602 3603
}

G
Glauber Costa 已提交
3604
int qemu_cpu_self(void *_env)
3605
{
G
Glauber Costa 已提交
3606 3607 3608 3609 3610 3611
    CPUState *env = _env;
    QemuThread this;
 
    qemu_thread_self(&this);
 
    return qemu_thread_equal(&this, env->thread);
3612 3613 3614 3615 3616 3617 3618 3619
}

static void cpu_signal(int sig)
{
    if (cpu_single_env)
        cpu_exit(cpu_single_env);
}

3620
static void tcg_block_io_signals(void)
3621 3622 3623 3624 3625 3626 3627 3628
{
    sigset_t set;
    struct sigaction sigact;

    sigemptyset(&set);
    sigaddset(&set, SIGUSR2);
    sigaddset(&set, SIGIO);
    sigaddset(&set, SIGALRM);
3629
    sigaddset(&set, SIGCHLD);
3630 3631 3632
    pthread_sigmask(SIG_BLOCK, &set, NULL);

    sigemptyset(&set);
3633
    sigaddset(&set, SIG_IPI);
3634 3635 3636 3637
    pthread_sigmask(SIG_UNBLOCK, &set, NULL);

    memset(&sigact, 0, sizeof(sigact));
    sigact.sa_handler = cpu_signal;
3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670
    sigaction(SIG_IPI, &sigact, NULL);
}

static void dummy_signal(int sig)
{
}

static void kvm_block_io_signals(CPUState *env)
{
    int r;
    sigset_t set;
    struct sigaction sigact;

    sigemptyset(&set);
    sigaddset(&set, SIGUSR2);
    sigaddset(&set, SIGIO);
    sigaddset(&set, SIGALRM);
    sigaddset(&set, SIGCHLD);
    sigaddset(&set, SIG_IPI);
    pthread_sigmask(SIG_BLOCK, &set, NULL);

    pthread_sigmask(SIG_BLOCK, NULL, &set);
    sigdelset(&set, SIG_IPI);

    memset(&sigact, 0, sizeof(sigact));
    sigact.sa_handler = dummy_signal;
    sigaction(SIG_IPI, &sigact, NULL);

    r = kvm_set_signal_mask(env, &set);
    if (r) {
        fprintf(stderr, "kvm_set_signal_mask: %s\n", strerror(r));
        exit(1);
    }
3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683
}

static void unblock_io_signals(void)
{
    sigset_t set;

    sigemptyset(&set);
    sigaddset(&set, SIGUSR2);
    sigaddset(&set, SIGIO);
    sigaddset(&set, SIGALRM);
    pthread_sigmask(SIG_UNBLOCK, &set, NULL);

    sigemptyset(&set);
3684
    sigaddset(&set, SIG_IPI);
3685 3686 3687 3688 3689 3690 3691 3692
    pthread_sigmask(SIG_BLOCK, &set, NULL);
}

static void qemu_signal_lock(unsigned int msecs)
{
    qemu_mutex_lock(&qemu_fair_mutex);

    while (qemu_mutex_trylock(&qemu_global_mutex)) {
3693
        qemu_thread_signal(tcg_cpu_thread, SIG_IPI);
3694 3695 3696 3697 3698 3699
        if (!qemu_mutex_timedlock(&qemu_global_mutex, msecs))
            break;
    }
    qemu_mutex_unlock(&qemu_fair_mutex);
}

3700
void qemu_mutex_lock_iothread(void)
3701 3702 3703 3704 3705 3706 3707 3708 3709
{
    if (kvm_enabled()) {
        qemu_mutex_lock(&qemu_fair_mutex);
        qemu_mutex_lock(&qemu_global_mutex);
        qemu_mutex_unlock(&qemu_fair_mutex);
    } else
        qemu_signal_lock(100);
}

3710
void qemu_mutex_unlock_iothread(void)
3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733
{
    qemu_mutex_unlock(&qemu_global_mutex);
}

static int all_vcpus_paused(void)
{
    CPUState *penv = first_cpu;

    while (penv) {
        if (!penv->stopped)
            return 0;
        penv = (CPUState *)penv->next_cpu;
    }

    return 1;
}

static void pause_all_vcpus(void)
{
    CPUState *penv = first_cpu;

    while (penv) {
        penv->stop = 1;
3734
        qemu_thread_signal(penv->thread, SIG_IPI);
3735 3736 3737 3738 3739 3740 3741 3742
        qemu_cpu_kick(penv);
        penv = (CPUState *)penv->next_cpu;
    }

    while (!all_vcpus_paused()) {
        qemu_cond_timedwait(&qemu_pause_cond, &qemu_global_mutex, 100);
        penv = first_cpu;
        while (penv) {
3743
            qemu_thread_signal(penv->thread, SIG_IPI);
3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755
            penv = (CPUState *)penv->next_cpu;
        }
    }
}

static void resume_all_vcpus(void)
{
    CPUState *penv = first_cpu;

    while (penv) {
        penv->stop = 0;
        penv->stopped = 0;
3756
        qemu_thread_signal(penv->thread, SIG_IPI);
3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794
        qemu_cpu_kick(penv);
        penv = (CPUState *)penv->next_cpu;
    }
}

static void tcg_init_vcpu(void *_env)
{
    CPUState *env = _env;
    /* share a single thread for all cpus with TCG */
    if (!tcg_cpu_thread) {
        env->thread = qemu_mallocz(sizeof(QemuThread));
        env->halt_cond = qemu_mallocz(sizeof(QemuCond));
        qemu_cond_init(env->halt_cond);
        qemu_thread_create(env->thread, tcg_cpu_thread_fn, env);
        while (env->created == 0)
            qemu_cond_timedwait(&qemu_cpu_cond, &qemu_global_mutex, 100);
        tcg_cpu_thread = env->thread;
        tcg_halt_cond = env->halt_cond;
    } else {
        env->thread = tcg_cpu_thread;
        env->halt_cond = tcg_halt_cond;
    }
}

static void kvm_start_vcpu(CPUState *env)
{
    env->thread = qemu_mallocz(sizeof(QemuThread));
    env->halt_cond = qemu_mallocz(sizeof(QemuCond));
    qemu_cond_init(env->halt_cond);
    qemu_thread_create(env->thread, kvm_cpu_thread_fn, env);
    while (env->created == 0)
        qemu_cond_timedwait(&qemu_cpu_cond, &qemu_global_mutex, 100);
}

void qemu_init_vcpu(void *_env)
{
    CPUState *env = _env;

J
Jiri Denemark 已提交
3795 3796
    env->nr_cores = smp_cores;
    env->nr_threads = smp_threads;
3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830
    if (kvm_enabled())
        kvm_start_vcpu(env);
    else
        tcg_init_vcpu(env);
}

void qemu_notify_event(void)
{
    qemu_event_increment();
}

void vm_stop(int reason)
{
    QemuThread me;
    qemu_thread_self(&me);

    if (!qemu_thread_equal(&me, &io_thread)) {
        qemu_system_vmstop_request(reason);
        /*
         * FIXME: should not return to device code in case
         * vm_stop() has been requested.
         */
        if (cpu_single_env) {
            cpu_exit(cpu_single_env);
            cpu_single_env->stop = 1;
        }
        return;
    }
    do_vm_stop(reason);
}

#endif


T
ths 已提交
3831
#ifdef _WIN32
3832
static void host_main_loop_wait(int *timeout)
A
aliguori 已提交
3833 3834
{
    int ret, ret2, i;
3835 3836
    PollingEntry *pe;

3837

3838 3839 3840 3841 3842
    /* XXX: need to suppress polling by better using win32 events */
    ret = 0;
    for(pe = first_polling_entry; pe != NULL; pe = pe->next) {
        ret |= pe->func(pe->opaque);
    }
3843
    if (ret == 0) {
3844 3845
        int err;
        WaitObjects *w = &wait_objects;
3846

A
aliguori 已提交
3847
        ret = WaitForMultipleObjects(w->num, w->events, FALSE, *timeout);
3848 3849 3850
        if (WAIT_OBJECT_0 + 0 <= ret && ret <= WAIT_OBJECT_0 + w->num - 1) {
            if (w->func[ret - WAIT_OBJECT_0])
                w->func[ret - WAIT_OBJECT_0](w->opaque[ret - WAIT_OBJECT_0]);
3851

3852
            /* Check for additional signaled events */
3853
            for(i = (ret - WAIT_OBJECT_0 + 1); i < w->num; i++) {
3854

3855 3856 3857 3858 3859 3860 3861 3862 3863
                /* Check if event is signaled */
                ret2 = WaitForSingleObject(w->events[i], 0);
                if(ret2 == WAIT_OBJECT_0) {
                    if (w->func[i])
                        w->func[i](w->opaque[i]);
                } else if (ret2 == WAIT_TIMEOUT) {
                } else {
                    err = GetLastError();
                    fprintf(stderr, "WaitForSingleObject error %d %d\n", i, err);
3864 3865
                }
            }
3866 3867 3868
        } else if (ret == WAIT_TIMEOUT) {
        } else {
            err = GetLastError();
3869
            fprintf(stderr, "WaitForMultipleObjects error %d %d\n", ret, err);
3870
        }
3871
    }
A
aliguori 已提交
3872 3873 3874 3875

    *timeout = 0;
}
#else
3876
static void host_main_loop_wait(int *timeout)
A
aliguori 已提交
3877 3878
{
}
B
bellard 已提交
3879
#endif
A
aliguori 已提交
3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891

void main_loop_wait(int timeout)
{
    IOHandlerRecord *ioh;
    fd_set rfds, wfds, xfds;
    int ret, nfds;
    struct timeval tv;

    qemu_bh_update_timeout(&timeout);

    host_main_loop_wait(&timeout);

B
bellard 已提交
3892 3893
    /* poll any events */
    /* XXX: separate device handlers from system ones */
3894
    nfds = -1;
B
bellard 已提交
3895 3896
    FD_ZERO(&rfds);
    FD_ZERO(&wfds);
3897
    FD_ZERO(&xfds);
B
bellard 已提交
3898
    for(ioh = first_io_handler; ioh != NULL; ioh = ioh->next) {
3899 3900
        if (ioh->deleted)
            continue;
B
bellard 已提交
3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913
        if (ioh->fd_read &&
            (!ioh->fd_read_poll ||
             ioh->fd_read_poll(ioh->opaque) != 0)) {
            FD_SET(ioh->fd, &rfds);
            if (ioh->fd > nfds)
                nfds = ioh->fd;
        }
        if (ioh->fd_write) {
            FD_SET(ioh->fd, &wfds);
            if (ioh->fd > nfds)
                nfds = ioh->fd;
        }
    }
3914

A
aliguori 已提交
3915 3916 3917
    tv.tv_sec = timeout / 1000;
    tv.tv_usec = (timeout % 1000) * 1000;

J
Jan Kiszka 已提交
3918 3919
    slirp_select_fill(&nfds, &rfds, &wfds, &xfds);

3920
    qemu_mutex_unlock_iothread();
3921
    ret = select(nfds + 1, &rfds, &wfds, &xfds, &tv);
3922
    qemu_mutex_lock_iothread();
B
bellard 已提交
3923
    if (ret > 0) {
3924 3925 3926
        IOHandlerRecord **pioh;

        for(ioh = first_io_handler; ioh != NULL; ioh = ioh->next) {
3927
            if (!ioh->deleted && ioh->fd_read && FD_ISSET(ioh->fd, &rfds)) {
B
bellard 已提交
3928
                ioh->fd_read(ioh->opaque);
B
bellard 已提交
3929
            }
3930
            if (!ioh->deleted && ioh->fd_write && FD_ISSET(ioh->fd, &wfds)) {
B
bellard 已提交
3931
                ioh->fd_write(ioh->opaque);
3932
            }
B
bellard 已提交
3933
        }
3934 3935 3936 3937 3938 3939 3940 3941

	/* remove deleted IO handlers */
	pioh = &first_io_handler;
	while (*pioh) {
            ioh = *pioh;
            if (ioh->deleted) {
                *pioh = ioh->next;
                qemu_free(ioh);
3942
            } else
3943 3944
                pioh = &ioh->next;
        }
B
bellard 已提交
3945
    }
J
Jan Kiszka 已提交
3946 3947

    slirp_select_poll(&rfds, &wfds, &xfds, (ret < 0));
B
bellard 已提交
3948

3949
    qemu_run_all_timers();
J
Jan Kiszka 已提交
3950

P
pbrook 已提交
3951 3952 3953
    /* Check bottom-halves last in case any of the earlier events triggered
       them.  */
    qemu_bh_poll();
3954

B
bellard 已提交
3955 3956
}

3957
static int qemu_cpu_exec(CPUState *env)
B
bellard 已提交
3958
{
3959
    int ret;
3960 3961 3962
#ifdef CONFIG_PROFILER
    int64_t ti;
#endif
B
bellard 已提交
3963

3964
#ifdef CONFIG_PROFILER
3965
    ti = profile_getclock();
3966
#endif
3967 3968 3969 3970 3971 3972
    if (use_icount) {
        int64_t count;
        int decr;
        qemu_icount -= (env->icount_decr.u16.low + env->icount_extra);
        env->icount_decr.u16.low = 0;
        env->icount_extra = 0;
P
Paolo Bonzini 已提交
3973
        count = qemu_icount_round (qemu_next_deadline());
3974 3975 3976 3977 3978 3979 3980
        qemu_icount += count;
        decr = (count > 0xffff) ? 0xffff : count;
        count -= decr;
        env->icount_decr.u16.low = decr;
        env->icount_extra = count;
    }
    ret = cpu_exec(env);
3981
#ifdef CONFIG_PROFILER
3982
    qemu_time += profile_getclock() - ti;
3983
#endif
3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994
    if (use_icount) {
        /* Fold pending instructions back into the
           instruction counter, and clear the interrupt flag.  */
        qemu_icount -= (env->icount_decr.u16.low
                        + env->icount_extra);
        env->icount_decr.u32 = 0;
        env->icount_extra = 0;
    }
    return ret;
}

3995 3996
static void tcg_cpu_exec(void)
{
3997
    int ret = 0;
3998 3999 4000 4001 4002 4003

    if (next_cpu == NULL)
        next_cpu = first_cpu;
    for (; next_cpu != NULL; next_cpu = next_cpu->next_cpu) {
        CPUState *env = cur_cpu = next_cpu;

4004 4005 4006
        qemu_clock_enable(vm_clock,
                          (cur_cpu->singlestep_enabled & SSTEP_NOTIMER) == 0);

P
Paolo Bonzini 已提交
4007
        if (qemu_alarm_pending())
4008
            break;
4009 4010
        if (cpu_can_run(env))
            ret = qemu_cpu_exec(env);
4011 4012 4013
        else if (env->stop)
            break;

4014 4015 4016 4017 4018 4019 4020 4021
        if (ret == EXCP_DEBUG) {
            gdb_set_stop_cpu(env);
            debug_requested = 1;
            break;
        }
    }
}

4022 4023
static int cpu_has_work(CPUState *env)
{
4024 4025 4026 4027
    if (env->stop)
        return 1;
    if (env->stopped)
        return 0;
4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046
    if (!env->halted)
        return 1;
    if (qemu_cpu_has_work(env))
        return 1;
    return 0;
}

static int tcg_has_work(void)
{
    CPUState *env;

    for (env = first_cpu; env != NULL; env = env->next_cpu)
        if (cpu_has_work(env))
            return 1;
    return 0;
}

static int qemu_calculate_timeout(void)
{
4047
#ifndef CONFIG_IOTHREAD
4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082
    int timeout;

    if (!vm_running)
        timeout = 5000;
    else if (tcg_has_work())
        timeout = 0;
    else if (!use_icount)
        timeout = 5000;
    else {
     /* XXX: use timeout computed from timers */
        int64_t add;
        int64_t delta;
        /* Advance virtual time to the next event.  */
        if (use_icount == 1) {
            /* When not using an adaptive execution frequency
               we tend to get badly out of sync with real time,
               so just delay for a reasonable amount of time.  */
            delta = 0;
        } else {
            delta = cpu_get_icount() - cpu_get_clock();
        }
        if (delta > 0) {
            /* If virtual time is ahead of real time then just
               wait for IO.  */
            timeout = (delta / 1000000) + 1;
        } else {
            /* Wait for either IO to occur or the next
               timer event.  */
            add = qemu_next_deadline();
            /* We advance the timer before checking for IO.
               Limit the amount we advance so that early IO
               activity won't get the guest too far ahead.  */
            if (add > 10000000)
                add = 10000000;
            delta += add;
P
Paolo Bonzini 已提交
4083
            qemu_icount += qemu_icount_round (add);
4084 4085 4086 4087 4088 4089 4090
            timeout = delta / 1000000;
            if (timeout < 0)
                timeout = 0;
        }
    }

    return timeout;
4091 4092 4093
#else /* CONFIG_IOTHREAD */
    return 1000;
#endif
4094 4095 4096 4097 4098 4099 4100 4101 4102 4103
}

static int vm_can_run(void)
{
    if (powerdown_requested)
        return 0;
    if (reset_requested)
        return 0;
    if (shutdown_requested)
        return 0;
4104 4105
    if (debug_requested)
        return 0;
4106 4107 4108
    return 1;
}

B
Blue Swirl 已提交
4109 4110
qemu_irq qemu_system_powerdown;

4111 4112
static void main_loop(void)
{
4113
    int r;
4114

4115 4116 4117 4118 4119
#ifdef CONFIG_IOTHREAD
    qemu_system_ready = 1;
    qemu_cond_broadcast(&qemu_system_cond);
#endif

4120
    for (;;) {
4121
        do {
4122 4123 4124
#ifdef CONFIG_PROFILER
            int64_t ti;
#endif
4125
#ifndef CONFIG_IOTHREAD
4126
            tcg_cpu_exec();
4127
#endif
4128
#ifdef CONFIG_PROFILER
4129
            ti = profile_getclock();
4130
#endif
4131
            main_loop_wait(qemu_calculate_timeout());
4132
#ifdef CONFIG_PROFILER
4133
            dev_time += profile_getclock() - ti;
4134
#endif
4135
        } while (vm_can_run());
4136

4137
        if (qemu_debug_requested()) {
4138
            vm_stop(EXCP_DEBUG);
4139
        }
4140
        if (qemu_shutdown_requested()) {
4141
            monitor_protocol_event(QEVENT_SHUTDOWN, NULL);
4142 4143 4144 4145 4146 4147
            if (no_shutdown) {
                vm_stop(0);
                no_shutdown = 0;
            } else
                break;
        }
4148 4149
        if (qemu_reset_requested()) {
            pause_all_vcpus();
4150
            qemu_system_reset();
4151 4152
            resume_all_vcpus();
        }
B
Blue Swirl 已提交
4153
        if (qemu_powerdown_requested()) {
4154
            monitor_protocol_event(QEVENT_POWERDOWN, NULL);
B
Blue Swirl 已提交
4155 4156
            qemu_irq_raise(qemu_system_powerdown);
        }
4157
        if ((r = qemu_vmstop_requested())) {
4158
            vm_stop(r);
4159
        }
B
bellard 已提交
4160
    }
4161
    pause_all_vcpus();
B
bellard 已提交
4162 4163
}

P
pbrook 已提交
4164 4165
static void version(void)
{
P
pbrook 已提交
4166
    printf("QEMU PC emulator version " QEMU_VERSION QEMU_PKGVERSION ", Copyright (c) 2003-2008 Fabrice Bellard\n");
P
pbrook 已提交
4167 4168
}

4169
static void help(int exitcode)
4170
{
4171
    const char *options_help =
4172 4173 4174 4175 4176 4177 4178
#define DEF(option, opt_arg, opt_enum, opt_help)        \
           opt_help
#define DEFHEADING(text) stringify(text) "\n"
#include "qemu-options.h"
#undef DEF
#undef DEFHEADING
#undef GEN_DOCS
4179 4180 4181 4182 4183
        ;
    version();
    printf("usage: %s [options] [disk_image]\n"
           "\n"
           "'disk_image' is a raw hard image image for IDE hard disk 0\n"
4184
           "\n"
4185
           "%s\n"
4186 4187 4188 4189 4190
           "During emulation, the following keys are useful:\n"
           "ctrl-alt-f      toggle full screen\n"
           "ctrl-alt-n      switch to virtual console 'n'\n"
           "ctrl-alt        toggle mouse and keyboard grab\n"
           "\n"
4191 4192 4193
           "When using -nographic, press 'ctrl-a h' to get some help.\n",
           "qemu",
           options_help);
4194
    exit(exitcode);
4195 4196
}

4197 4198 4199
#define HAS_ARG 0x0001

enum {
4200 4201 4202 4203 4204 4205 4206
#define DEF(option, opt_arg, opt_enum, opt_help)        \
    opt_enum,
#define DEFHEADING(text)
#include "qemu-options.h"
#undef DEF
#undef DEFHEADING
#undef GEN_DOCS
4207 4208 4209 4210 4211 4212 4213 4214
};

typedef struct QEMUOption {
    const char *name;
    int flags;
    int index;
} QEMUOption;

B
blueswir1 已提交
4215
static const QEMUOption qemu_options[] = {
4216
    { "h", 0, QEMU_OPTION_h },
4217 4218 4219 4220 4221 4222 4223
#define DEF(option, opt_arg, opt_enum, opt_help)        \
    { option, opt_arg, opt_enum },
#define DEFHEADING(text)
#include "qemu-options.h"
#undef DEF
#undef DEFHEADING
#undef GEN_DOCS
4224
    { NULL },
B
bellard 已提交
4225 4226
};

4227
#ifdef HAS_AUDIO
4228
struct soundhw soundhw[] = {
4229
#ifdef HAS_AUDIO_CHOICE
A
aurel32 已提交
4230
#if defined(TARGET_I386) || defined(TARGET_MIPS)
B
bellard 已提交
4231 4232 4233 4234 4235 4236 4237 4238
    {
        "pcspk",
        "PC speaker",
        0,
        1,
        { .init_isa = pcspk_audio_init }
    },
#endif
M
malc 已提交
4239 4240

#ifdef CONFIG_SB16
4241 4242 4243 4244 4245 4246 4247
    {
        "sb16",
        "Creative Sound Blaster 16",
        0,
        1,
        { .init_isa = SB16_init }
    },
M
malc 已提交
4248
#endif
4249

M
malc 已提交
4250 4251 4252 4253 4254 4255 4256 4257 4258 4259
#ifdef CONFIG_CS4231A
    {
        "cs4231a",
        "CS4231A",
        0,
        1,
        { .init_isa = cs4231a_init }
    },
#endif

4260
#ifdef CONFIG_ADLIB
4261 4262
    {
        "adlib",
4263
#ifdef HAS_YMF262
4264
        "Yamaha YMF262 (OPL3)",
4265
#else
4266
        "Yamaha YM3812 (OPL2)",
4267
#endif
4268 4269 4270 4271
        0,
        1,
        { .init_isa = Adlib_init }
    },
4272
#endif
4273

4274
#ifdef CONFIG_GUS
4275 4276 4277 4278 4279 4280 4281
    {
        "gus",
        "Gravis Ultrasound GF1",
        0,
        1,
        { .init_isa = GUS_init }
    },
4282
#endif
4283

M
malc 已提交
4284
#ifdef CONFIG_AC97
B
balrog 已提交
4285 4286 4287 4288 4289 4290 4291
    {
        "ac97",
        "Intel 82801AA AC97 Audio",
        0,
        0,
        { .init_pci = ac97_init }
    },
M
malc 已提交
4292
#endif
B
balrog 已提交
4293

M
malc 已提交
4294
#ifdef CONFIG_ES1370
4295 4296 4297 4298 4299 4300 4301
    {
        "es1370",
        "ENSONIQ AudioPCI ES1370",
        0,
        0,
        { .init_pci = es1370_init }
    },
4302
#endif
4303

M
malc 已提交
4304 4305
#endif /* HAS_AUDIO_CHOICE */

4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320
    { NULL, NULL, 0, 0, { NULL } }
};

static void select_soundhw (const char *optarg)
{
    struct soundhw *c;

    if (*optarg == '?') {
    show_valid_cards:

        printf ("Valid sound card names (comma separated):\n");
        for (c = soundhw; c->name; ++c) {
            printf ("%-11s %s\n", c->name, c->descr);
        }
        printf ("\n-soundhw all will enable all of the above\n");
4321 4322 4323
        exit (*optarg != '?');
    }
    else {
4324
        size_t l;
4325 4326 4327 4328
        const char *p;
        char *e;
        int bad_card = 0;

4329 4330 4331 4332 4333 4334
        if (!strcmp (optarg, "all")) {
            for (c = soundhw; c->name; ++c) {
                c->enabled = 1;
            }
            return;
        }
4335

4336
        p = optarg;
4337 4338 4339
        while (*p) {
            e = strchr (p, ',');
            l = !e ? strlen (p) : (size_t) (e - p);
4340 4341

            for (c = soundhw; c->name; ++c) {
4342
                if (!strncmp (c->name, p, l) && !c->name[l]) {
4343
                    c->enabled = 1;
4344 4345 4346
                    break;
                }
            }
4347 4348

            if (!c->name) {
4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367
                if (l > 80) {
                    fprintf (stderr,
                             "Unknown sound card name (too big to show)\n");
                }
                else {
                    fprintf (stderr, "Unknown sound card name `%.*s'\n",
                             (int) l, p);
                }
                bad_card = 1;
            }
            p += l + (e != NULL);
        }

        if (bad_card)
            goto show_valid_cards;
    }
}
#endif

4368 4369 4370 4371
static void select_vgahw (const char *p)
{
    const char *opts;

G
Gerd Hoffmann 已提交
4372
    default_vga = 0;
4373
    vga_interface_type = VGA_NONE;
4374
    if (strstart(p, "std", &opts)) {
4375
        vga_interface_type = VGA_STD;
4376
    } else if (strstart(p, "cirrus", &opts)) {
4377
        vga_interface_type = VGA_CIRRUS;
4378
    } else if (strstart(p, "vmware", &opts)) {
4379
        vga_interface_type = VGA_VMWARE;
4380
    } else if (strstart(p, "xenfb", &opts)) {
4381
        vga_interface_type = VGA_XENFB;
4382
    } else if (!strstart(p, "none", &opts)) {
4383 4384 4385 4386
    invalid_vga:
        fprintf(stderr, "Unknown vga type: %s\n", p);
        exit(1);
    }
M
malc 已提交
4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399
    while (*opts) {
        const char *nextopt;

        if (strstart(opts, ",retrace=", &nextopt)) {
            opts = nextopt;
            if (strstart(opts, "dumb", &nextopt))
                vga_retrace_method = VGA_RETRACE_DUMB;
            else if (strstart(opts, "precise", &nextopt))
                vga_retrace_method = VGA_RETRACE_PRECISE;
            else goto invalid_vga;
        } else goto invalid_vga;
        opts = nextopt;
    }
4400 4401
}

4402 4403 4404
#ifdef TARGET_I386
static int balloon_parse(const char *arg)
{
4405
    QemuOpts *opts;
4406

4407 4408 4409 4410 4411 4412 4413
    if (strcmp(arg, "none") == 0) {
        return 0;
    }

    if (!strncmp(arg, "virtio", 6)) {
        if (arg[6] == ',') {
            /* have params -> parse them */
4414
            opts = qemu_opts_parse(&qemu_device_opts, arg+7, 0);
4415 4416 4417 4418 4419
            if (!opts)
                return  -1;
        } else {
            /* create empty opts */
            opts = qemu_opts_create(&qemu_device_opts, NULL, 0);
4420
        }
4421 4422
        qemu_opt_set(opts, "driver", "virtio-balloon-pci");
        return 0;
4423
    }
4424 4425

    return -1;
4426 4427 4428
}
#endif

4429 4430 4431 4432 4433 4434 4435 4436
#ifdef _WIN32
static BOOL WINAPI qemu_ctrl_handler(DWORD type)
{
    exit(STATUS_CONTROL_C_EXIT);
    return TRUE;
}
#endif

4437
int qemu_uuid_parse(const char *str, uint8_t *uuid)
4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450
{
    int ret;

    if(strlen(str) != 36)
        return -1;

    ret = sscanf(str, UUID_FMT, &uuid[0], &uuid[1], &uuid[2], &uuid[3],
            &uuid[4], &uuid[5], &uuid[6], &uuid[7], &uuid[8], &uuid[9],
            &uuid[10], &uuid[11], &uuid[12], &uuid[13], &uuid[14], &uuid[15]);

    if(ret != 16)
        return -1;

4451 4452 4453 4454
#ifdef TARGET_I386
    smbios_add_field(1, offsetof(struct smbios_type_1, uuid), 16, uuid);
#endif

4455 4456 4457
    return 0;
}

4458 4459 4460 4461 4462 4463 4464
#ifndef _WIN32

static void termsig_handler(int signal)
{
    qemu_system_shutdown_request();
}

4465 4466 4467 4468 4469 4470
static void sigchld_handler(int signal)
{
    waitpid(-1, NULL, WNOHANG);
}

static void sighandler_setup(void)
4471 4472 4473 4474 4475 4476 4477 4478
{
    struct sigaction act;

    memset(&act, 0, sizeof(act));
    act.sa_handler = termsig_handler;
    sigaction(SIGINT,  &act, NULL);
    sigaction(SIGHUP,  &act, NULL);
    sigaction(SIGTERM, &act, NULL);
4479 4480 4481 4482

    act.sa_handler = sigchld_handler;
    act.sa_flags = SA_NOCLDSTOP;
    sigaction(SIGCHLD, &act, NULL);
4483 4484 4485 4486
}

#endif

P
Paul Brook 已提交
4487 4488 4489 4490 4491 4492 4493 4494 4495 4496
#ifdef _WIN32
/* Look for support files in the same directory as the executable.  */
static char *find_datadir(const char *argv0)
{
    char *p;
    char buf[MAX_PATH];
    DWORD len;

    len = GetModuleFileName(NULL, buf, sizeof(buf) - 1);
    if (len == 0) {
B
Blue Swirl 已提交
4497
        return NULL;
P
Paul Brook 已提交
4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522
    }

    buf[len] = 0;
    p = buf + len - 1;
    while (p != buf && *p != '\\')
        p--;
    *p = 0;
    if (access(buf, R_OK) == 0) {
        return qemu_strdup(buf);
    }
    return NULL;
}
#else /* !_WIN32 */

/* Find a likely location for support files using the location of the binary.
   For installed binaries this will be "$bindir/../share/qemu".  When
   running from the build tree this will be "$bindir/../pc-bios".  */
#define SHARE_SUFFIX "/share/qemu"
#define BUILD_SUFFIX "/pc-bios"
static char *find_datadir(const char *argv0)
{
    char *dir;
    char *p = NULL;
    char *res;
    char buf[PATH_MAX];
4523
    size_t max_len;
P
Paul Brook 已提交
4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546

#if defined(__linux__)
    {
        int len;
        len = readlink("/proc/self/exe", buf, sizeof(buf) - 1);
        if (len > 0) {
            buf[len] = 0;
            p = buf;
        }
    }
#elif defined(__FreeBSD__)
    {
        int len;
        len = readlink("/proc/curproc/file", buf, sizeof(buf) - 1);
        if (len > 0) {
            buf[len] = 0;
            p = buf;
        }
    }
#endif
    /* If we don't have any way of figuring out the actual executable
       location then try argv[0].  */
    if (!p) {
4547
        p = realpath(argv0, buf);
P
Paul Brook 已提交
4548 4549 4550 4551 4552 4553 4554
        if (!p) {
            return NULL;
        }
    }
    dir = dirname(p);
    dir = dirname(dir);

4555 4556 4557 4558
    max_len = strlen(dir) +
        MAX(strlen(SHARE_SUFFIX), strlen(BUILD_SUFFIX)) + 1;
    res = qemu_mallocz(max_len);
    snprintf(res, max_len, "%s%s", dir, SHARE_SUFFIX);
P
Paul Brook 已提交
4559
    if (access(res, R_OK)) {
4560
        snprintf(res, max_len, "%s%s", dir, BUILD_SUFFIX);
P
Paul Brook 已提交
4561 4562 4563 4564 4565
        if (access(res, R_OK)) {
            qemu_free(res);
            res = NULL;
        }
    }
4566

P
Paul Brook 已提交
4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581
    return res;
}
#undef SHARE_SUFFIX
#undef BUILD_SUFFIX
#endif

char *qemu_find_file(int type, const char *name)
{
    int len;
    const char *subdir;
    char *buf;

    /* If name contains path separators then try it as a straight path.  */
    if ((strchr(name, '/') || strchr(name, '\\'))
        && access(name, R_OK) == 0) {
4582
        return qemu_strdup(name);
P
Paul Brook 已提交
4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595
    }
    switch (type) {
    case QEMU_FILE_TYPE_BIOS:
        subdir = "";
        break;
    case QEMU_FILE_TYPE_KEYMAP:
        subdir = "keymaps/";
        break;
    default:
        abort();
    }
    len = strlen(data_dir) + strlen(name) + strlen(subdir) + 2;
    buf = qemu_mallocz(len);
4596
    snprintf(buf, len, "%s/%s%s", data_dir, subdir, name);
P
Paul Brook 已提交
4597 4598 4599 4600 4601 4602 4603
    if (access(buf, R_OK)) {
        qemu_free(buf);
        return NULL;
    }
    return buf;
}

4604 4605 4606 4607 4608
static int device_help_func(QemuOpts *opts, void *opaque)
{
    return qdev_device_help(opts);
}

G
Gerd Hoffmann 已提交
4609 4610 4611 4612 4613 4614 4615 4616 4617 4618
static int device_init_func(QemuOpts *opts, void *opaque)
{
    DeviceState *dev;

    dev = qdev_device_add(opts);
    if (!dev)
        return -1;
    return 0;
}

4619 4620 4621 4622 4623 4624 4625 4626 4627 4628
static int chardev_init_func(QemuOpts *opts, void *opaque)
{
    CharDriverState *chr;

    chr = qemu_chr_open_opts(opts, NULL);
    if (!chr)
        return -1;
    return 0;
}

4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662
static int mon_init_func(QemuOpts *opts, void *opaque)
{
    CharDriverState *chr;
    const char *chardev;
    const char *mode;
    int flags;

    mode = qemu_opt_get(opts, "mode");
    if (mode == NULL) {
        mode = "readline";
    }
    if (strcmp(mode, "readline") == 0) {
        flags = MONITOR_USE_READLINE;
    } else if (strcmp(mode, "control") == 0) {
        flags = MONITOR_USE_CONTROL;
    } else {
        fprintf(stderr, "unknown monitor mode \"%s\"\n", mode);
        exit(1);
    }

    if (qemu_opt_get_bool(opts, "default", 0))
        flags |= MONITOR_IS_DEFAULT;

    chardev = qemu_opt_get(opts, "chardev");
    chr = qemu_chr_find(chardev);
    if (chr == NULL) {
        fprintf(stderr, "chardev \"%s\" not found\n", chardev);
        exit(1);
    }

    monitor_init(chr, flags);
    return 0;
}

G
Gerd Hoffmann 已提交
4663
static void monitor_parse(const char *optarg, const char *mode)
4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692
{
    static int monitor_device_index = 0;
    QemuOpts *opts;
    const char *p;
    char label[32];
    int def = 0;

    if (strstart(optarg, "chardev:", &p)) {
        snprintf(label, sizeof(label), "%s", p);
    } else {
        if (monitor_device_index) {
            snprintf(label, sizeof(label), "monitor%d",
                     monitor_device_index);
        } else {
            snprintf(label, sizeof(label), "monitor");
            def = 1;
        }
        opts = qemu_chr_parse_compat(label, optarg);
        if (!opts) {
            fprintf(stderr, "parse error: %s\n", optarg);
            exit(1);
        }
    }

    opts = qemu_opts_create(&qemu_mon_opts, label, 1);
    if (!opts) {
        fprintf(stderr, "duplicate chardev: %s\n", label);
        exit(1);
    }
G
Gerd Hoffmann 已提交
4693
    qemu_opt_set(opts, "mode", mode);
4694 4695 4696 4697 4698 4699
    qemu_opt_set(opts, "chardev", label);
    if (def)
        qemu_opt_set(opts, "default", "on");
    monitor_device_index++;
}

4700 4701
struct device_config {
    enum {
G
Gerd Hoffmann 已提交
4702 4703 4704 4705 4706
        DEV_USB,       /* -usbdevice     */
        DEV_BT,        /* -bt            */
        DEV_SERIAL,    /* -serial        */
        DEV_PARALLEL,  /* -parallel      */
        DEV_VIRTCON,   /* -virtioconsole */
4707
        DEV_DEBUGCON,  /* -debugcon */
4708 4709
    } type;
    const char *cmdline;
B
Blue Swirl 已提交
4710
    QTAILQ_ENTRY(device_config) next;
4711
};
B
Blue Swirl 已提交
4712
QTAILQ_HEAD(, device_config) device_configs = QTAILQ_HEAD_INITIALIZER(device_configs);
4713 4714 4715 4716 4717 4718 4719 4720

static void add_device_config(int type, const char *cmdline)
{
    struct device_config *conf;

    conf = qemu_mallocz(sizeof(*conf));
    conf->type = type;
    conf->cmdline = cmdline;
B
Blue Swirl 已提交
4721
    QTAILQ_INSERT_TAIL(&device_configs, conf, next);
4722 4723 4724 4725 4726 4727 4728
}

static int foreach_device_config(int type, int (*func)(const char *cmdline))
{
    struct device_config *conf;
    int rc;

B
Blue Swirl 已提交
4729
    QTAILQ_FOREACH(conf, &device_configs, next) {
4730 4731 4732 4733 4734 4735 4736 4737 4738
        if (conf->type != type)
            continue;
        rc = func(conf->cmdline);
        if (0 != rc)
            return rc;
    }
    return 0;
}

4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760
static int serial_parse(const char *devname)
{
    static int index = 0;
    char label[32];

    if (strcmp(devname, "none") == 0)
        return 0;
    if (index == MAX_SERIAL_PORTS) {
        fprintf(stderr, "qemu: too many serial ports\n");
        exit(1);
    }
    snprintf(label, sizeof(label), "serial%d", index);
    serial_hds[index] = qemu_chr_open(label, devname, NULL);
    if (!serial_hds[index]) {
        fprintf(stderr, "qemu: could not open serial device '%s': %s\n",
                devname, strerror(errno));
        return -1;
    }
    index++;
    return 0;
}

G
Gerd Hoffmann 已提交
4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782
static int parallel_parse(const char *devname)
{
    static int index = 0;
    char label[32];

    if (strcmp(devname, "none") == 0)
        return 0;
    if (index == MAX_PARALLEL_PORTS) {
        fprintf(stderr, "qemu: too many parallel ports\n");
        exit(1);
    }
    snprintf(label, sizeof(label), "parallel%d", index);
    parallel_hds[index] = qemu_chr_open(label, devname, NULL);
    if (!parallel_hds[index]) {
        fprintf(stderr, "qemu: could not open parallel device '%s': %s\n",
                devname, strerror(errno));
        return -1;
    }
    index++;
    return 0;
}

G
Gerd Hoffmann 已提交
4783 4784 4785 4786
static int virtcon_parse(const char *devname)
{
    static int index = 0;
    char label[32];
4787
    QemuOpts *bus_opts, *dev_opts;
G
Gerd Hoffmann 已提交
4788 4789 4790 4791 4792 4793 4794

    if (strcmp(devname, "none") == 0)
        return 0;
    if (index == MAX_VIRTIO_CONSOLES) {
        fprintf(stderr, "qemu: too many virtio consoles\n");
        exit(1);
    }
4795 4796 4797 4798 4799 4800 4801

    bus_opts = qemu_opts_create(&qemu_device_opts, NULL, 0);
    qemu_opt_set(bus_opts, "driver", "virtio-serial");

    dev_opts = qemu_opts_create(&qemu_device_opts, NULL, 0);
    qemu_opt_set(dev_opts, "driver", "virtconsole");

G
Gerd Hoffmann 已提交
4802 4803 4804 4805 4806 4807 4808
    snprintf(label, sizeof(label), "virtcon%d", index);
    virtcon_hds[index] = qemu_chr_open(label, devname, NULL);
    if (!virtcon_hds[index]) {
        fprintf(stderr, "qemu: could not open virtio console '%s': %s\n",
                devname, strerror(errno));
        return -1;
    }
4809 4810
    qemu_opt_set(dev_opts, "chardev", label);

G
Gerd Hoffmann 已提交
4811 4812 4813 4814
    index++;
    return 0;
}

4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831
static int debugcon_parse(const char *devname)
{   
    QemuOpts *opts;

    if (!qemu_chr_open("debugcon", devname, NULL)) {
        exit(1);
    }
    opts = qemu_opts_create(&qemu_device_opts, "debugcon", 1);
    if (!opts) {
        fprintf(stderr, "qemu: already have a debugcon device\n");
        exit(1);
    }
    qemu_opt_set(opts, "driver", "isa-debugcon");
    qemu_opt_set(opts, "chardev", "debugcon");
    return 0;
}

4832 4833 4834 4835 4836 4837 4838 4839
static const QEMUOption *lookup_opt(int argc, char **argv,
                                    const char **poptarg, int *poptind)
{
    const QEMUOption *popt;
    int optind = *poptind;
    char *r = argv[optind];
    const char *optarg;

4840
    loc_set_cmdline(argv, optind, 1);
4841 4842 4843 4844 4845 4846 4847
    optind++;
    /* Treat --foo the same as -foo.  */
    if (r[1] == '-')
        r++;
    popt = qemu_options;
    for(;;) {
        if (!popt->name) {
4848
            error_report("invalid option");
4849 4850 4851 4852 4853 4854 4855 4856
            exit(1);
        }
        if (!strcmp(popt->name, r + 1))
            break;
        popt++;
    }
    if (popt->flags & HAS_ARG) {
        if (optind >= argc) {
4857
            error_report("requires an argument");
4858 4859 4860
            exit(1);
        }
        optarg = argv[optind++];
4861
        loc_set_cmdline(argv, optind - 2, 2);
4862 4863 4864 4865 4866 4867 4868 4869 4870 4871
    } else {
        optarg = NULL;
    }

    *poptarg = optarg;
    *poptind = optind;

    return popt;
}

M
malc 已提交
4872
int main(int argc, char **argv, char **envp)
4873
{
4874
    const char *gdbstub_dev = NULL;
4875
    uint32_t boot_devices_bitmap = 0;
T
ths 已提交
4876
    int i;
4877
    int snapshot, linux_boot, net_boot;
P
Paolo Bonzini 已提交
4878
    const char *icount_option = NULL;
B
bellard 已提交
4879
    const char *initrd_filename;
4880
    const char *kernel_filename, *kernel_cmdline;
4881
    char boot_devices[33] = "cad"; /* default to HD->floppy->CD-ROM */
4882
    DisplayState *ds;
4883
    DisplayChangeListener *dcl;
B
bellard 已提交
4884
    int cyls, heads, secs, translation;
G
Gerd Hoffmann 已提交
4885
    QemuOpts *hda_opts = NULL, *opts;
4886
    int optind;
4887
    const char *optarg;
B
bellard 已提交
4888
    const char *loadvm = NULL;
4889
    QEMUMachine *machine;
4890
    const char *cpu_model;
B
blueswir1 已提交
4891
#ifndef _WIN32
T
ths 已提交
4892
    int fds[2];
B
blueswir1 已提交
4893
#endif
4894
    int tb_size;
4895
    const char *pid_file = NULL;
A
aliguori 已提交
4896
    const char *incoming = NULL;
B
blueswir1 已提交
4897
#ifndef _WIN32
4898 4899
    int fd = 0;
    struct passwd *pwd = NULL;
4900 4901
    const char *chroot_dir = NULL;
    const char *run_as = NULL;
B
blueswir1 已提交
4902
#endif
4903
    CPUState *env;
4904
    int show_vnc_port = 0;
4905
    int defconfig = 1;
B
bellard 已提交
4906

4907 4908
    error_set_progname(argv[0]);

J
Jan Kiszka 已提交
4909 4910
    init_clocks();

M
malc 已提交
4911 4912
    qemu_cache_utils_init(envp);

B
Blue Swirl 已提交
4913
    QLIST_INIT (&vm_change_state_head);
4914 4915 4916 4917 4918 4919 4920 4921
#ifndef _WIN32
    {
        struct sigaction act;
        sigfillset(&act.sa_mask);
        act.sa_flags = 0;
        act.sa_handler = SIG_IGN;
        sigaction(SIGPIPE, &act, NULL);
    }
4922 4923
#else
    SetConsoleCtrlHandler(qemu_ctrl_handler, TRUE);
4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941
    /* Note: cpu_interrupt() is currently not SMP safe, so we force
       QEMU to run on a single CPU */
    {
        HANDLE h;
        DWORD mask, smask;
        int i;
        h = GetCurrentProcess();
        if (GetProcessAffinityMask(h, &mask, &smask)) {
            for(i = 0; i < 32; i++) {
                if (mask & (1 << i))
                    break;
            }
            if (i != 32) {
                mask = 1 << i;
                SetProcessAffinityMask(h, mask);
            }
        }
    }
B
bellard 已提交
4942
#endif
4943

4944
    module_call_init(MODULE_INIT_MACHINE);
4945
    machine = find_default_machine();
4946
    cpu_model = NULL;
B
bellard 已提交
4947
    initrd_filename = NULL;
4948
    ram_size = 0;
4949
    snapshot = 0;
4950 4951
    kernel_filename = NULL;
    kernel_cmdline = "";
4952
    cyls = heads = secs = 0;
B
bellard 已提交
4953
    translation = BIOS_ATA_TRANSLATION_AUTO;
4954

4955 4956 4957 4958 4959 4960
    for (i = 0; i < MAX_NODES; i++) {
        node_mem[i] = 0;
        node_cpumask[i] = 0;
    }

    nb_numa_nodes = 0;
B
bellard 已提交
4961
    nb_nics = 0;
4962

4963
    tb_size = 0;
4964 4965
    autostart= 1;

4966 4967 4968 4969 4970
    /* first pass of option parsing */
    optind = 1;
    while (optind < argc) {
        if (argv[optind][0] != '-') {
            /* disk image */
4971
            optind++;
4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985
            continue;
        } else {
            const QEMUOption *popt;

            popt = lookup_opt(argc, argv, &optarg, &optind);
            switch (popt->index) {
            case QEMU_OPTION_nodefconfig:
                defconfig=0;
                break;
            }
        }
    }

    if (defconfig) {
4986
        const char *fname;
4987
        FILE *fp;
4988 4989 4990

        fname = CONFIG_QEMU_CONFDIR "/qemu.conf";
        fp = fopen(fname, "r");
4991
        if (fp) {
4992
            if (qemu_config_parse(fp, fname) != 0) {
4993 4994 4995 4996 4997
                exit(1);
            }
            fclose(fp);
        }

4998 4999
        fname = CONFIG_QEMU_CONFDIR "/target-" TARGET_ARCH ".conf";
        fp = fopen(fname, "r");
5000
        if (fp) {
5001
            if (qemu_config_parse(fp, fname) != 0) {
5002 5003 5004 5005 5006
                exit(1);
            }
            fclose(fp);
        }
    }
5007 5008 5009
#if defined(cpudef_setup)
    cpudef_setup(); /* parse cpu definitions in target config file */
#endif
5010 5011

    /* second pass of option parsing */
5012
    optind = 1;
5013
    for(;;) {
5014
        if (optind >= argc)
5015
            break;
5016
        if (argv[optind][0] != '-') {
G
Gerd Hoffmann 已提交
5017
	    hda_opts = drive_add(argv[optind++], HD_ALIAS, 0);
5018 5019 5020
        } else {
            const QEMUOption *popt;

5021
            popt = lookup_opt(argc, argv, &optarg, &optind);
5022
            switch(popt->index) {
5023 5024 5025 5026 5027 5028
            case QEMU_OPTION_M:
                machine = find_machine(optarg);
                if (!machine) {
                    QEMUMachine *m;
                    printf("Supported machines are:\n");
                    for(m = first_machine; m != NULL; m = m->next) {
M
Mark McLoughlin 已提交
5029 5030 5031
                        if (m->alias)
                            printf("%-10s %s (alias of %s)\n",
                                   m->alias, m->desc, m->name);
5032
                        printf("%-10s %s%s\n",
5033
                               m->name, m->desc,
5034
                               m->is_default ? " (default)" : "");
5035
                    }
5036
                    exit(*optarg != '?');
5037 5038
                }
                break;
5039 5040
            case QEMU_OPTION_cpu:
                /* hw initialization will check this */
5041
                if (*optarg == '?') {
J
j_mayer 已提交
5042
/* XXX: implement xxx_cpu_list for targets that still miss it */
5043 5044 5045 5046
#if defined(cpu_list_id)
                    cpu_list_id(stdout, &fprintf, optarg);
#elif defined(cpu_list)
                    cpu_list(stdout, &fprintf);	        /* deprecated */
5047
#endif
5048
                    exit(0);
5049 5050 5051 5052
                } else {
                    cpu_model = optarg;
                }
                break;
5053
            case QEMU_OPTION_initrd:
B
bellard 已提交
5054 5055
                initrd_filename = optarg;
                break;
5056
            case QEMU_OPTION_hda:
T
ths 已提交
5057
                if (cyls == 0)
G
Gerd Hoffmann 已提交
5058
                    hda_opts = drive_add(optarg, HD_ALIAS, 0);
T
ths 已提交
5059
                else
G
Gerd Hoffmann 已提交
5060
                    hda_opts = drive_add(optarg, HD_ALIAS
T
ths 已提交
5061
			     ",cyls=%d,heads=%d,secs=%d%s",
5062
                             0, cyls, heads, secs,
T
ths 已提交
5063 5064 5065 5066 5067
                             translation == BIOS_ATA_TRANSLATION_LBA ?
                                 ",trans=lba" :
                             translation == BIOS_ATA_TRANSLATION_NONE ?
                                 ",trans=none" : "");
                 break;
5068
            case QEMU_OPTION_hdb:
5069 5070
            case QEMU_OPTION_hdc:
            case QEMU_OPTION_hdd:
5071
                drive_add(optarg, HD_ALIAS, popt->index - QEMU_OPTION_hda);
B
bellard 已提交
5072
                break;
T
ths 已提交
5073
            case QEMU_OPTION_drive:
5074
                drive_add(NULL, "%s", optarg);
T
ths 已提交
5075
	        break;
G
Gerd Hoffmann 已提交
5076 5077 5078 5079
            case QEMU_OPTION_set:
                if (qemu_set_option(optarg) != 0)
                    exit(1);
	        break;
5080 5081 5082 5083
            case QEMU_OPTION_global:
                if (qemu_global_option(optarg) != 0)
                    exit(1);
	        break;
5084
            case QEMU_OPTION_mtdblock:
5085
                drive_add(optarg, MTD_ALIAS);
5086
                break;
5087
            case QEMU_OPTION_sd:
5088
                drive_add(optarg, SD_ALIAS);
5089
                break;
5090
            case QEMU_OPTION_pflash:
5091
                drive_add(optarg, PFLASH_ALIAS);
5092
                break;
5093
            case QEMU_OPTION_snapshot:
5094 5095
                snapshot = 1;
                break;
5096
            case QEMU_OPTION_hdachs:
5097 5098 5099 5100
                {
                    const char *p;
                    p = optarg;
                    cyls = strtol(p, (char **)&p, 0);
B
bellard 已提交
5101 5102
                    if (cyls < 1 || cyls > 16383)
                        goto chs_fail;
5103 5104 5105 5106
                    if (*p != ',')
                        goto chs_fail;
                    p++;
                    heads = strtol(p, (char **)&p, 0);
B
bellard 已提交
5107 5108
                    if (heads < 1 || heads > 16)
                        goto chs_fail;
5109 5110 5111 5112
                    if (*p != ',')
                        goto chs_fail;
                    p++;
                    secs = strtol(p, (char **)&p, 0);
B
bellard 已提交
5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125
                    if (secs < 1 || secs > 63)
                        goto chs_fail;
                    if (*p == ',') {
                        p++;
                        if (!strcmp(p, "none"))
                            translation = BIOS_ATA_TRANSLATION_NONE;
                        else if (!strcmp(p, "lba"))
                            translation = BIOS_ATA_TRANSLATION_LBA;
                        else if (!strcmp(p, "auto"))
                            translation = BIOS_ATA_TRANSLATION_AUTO;
                        else
                            goto chs_fail;
                    } else if (*p != '\0') {
5126
                    chs_fail:
B
bellard 已提交
5127 5128
                        fprintf(stderr, "qemu: invalid physical CHS format\n");
                        exit(1);
5129
                    }
G
Gerd Hoffmann 已提交
5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142
		    if (hda_opts != NULL) {
                        char num[16];
                        snprintf(num, sizeof(num), "%d", cyls);
                        qemu_opt_set(hda_opts, "cyls", num);
                        snprintf(num, sizeof(num), "%d", heads);
                        qemu_opt_set(hda_opts, "heads", num);
                        snprintf(num, sizeof(num), "%d", secs);
                        qemu_opt_set(hda_opts, "secs", num);
                        if (translation == BIOS_ATA_TRANSLATION_LBA)
                            qemu_opt_set(hda_opts, "trans", "lba");
                        if (translation == BIOS_ATA_TRANSLATION_NONE)
                            qemu_opt_set(hda_opts, "trans", "none");
                    }
5143 5144
                }
                break;
5145 5146 5147 5148 5149 5150 5151
            case QEMU_OPTION_numa:
                if (nb_numa_nodes >= MAX_NODES) {
                    fprintf(stderr, "qemu: too many NUMA nodes\n");
                    exit(1);
                }
                numa_add(optarg);
                break;
5152
            case QEMU_OPTION_nographic:
5153
                display_type = DT_NOGRAPHIC;
5154
                break;
B
balrog 已提交
5155 5156
#ifdef CONFIG_CURSES
            case QEMU_OPTION_curses:
5157
                display_type = DT_CURSES;
B
balrog 已提交
5158 5159
                break;
#endif
5160 5161 5162
            case QEMU_OPTION_portrait:
                graphic_rotate = 1;
                break;
5163
            case QEMU_OPTION_kernel:
5164 5165
                kernel_filename = optarg;
                break;
5166
            case QEMU_OPTION_append:
5167
                kernel_cmdline = optarg;
5168
                break;
5169
            case QEMU_OPTION_cdrom:
5170
                drive_add(optarg, CDROM_ALIAS);
5171
                break;
5172
            case QEMU_OPTION_boot:
5173
                {
J
Jan Kiszka 已提交
5174
                    static const char * const params[] = {
5175
                        "order", "once", "menu", NULL
J
Jan Kiszka 已提交
5176 5177
                    };
                    char buf[sizeof(boot_devices)];
J
Jan Kiszka 已提交
5178
                    char *standard_boot_devices;
J
Jan Kiszka 已提交
5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194
                    int legacy = 0;

                    if (!strchr(optarg, '=')) {
                        legacy = 1;
                        pstrcpy(buf, sizeof(buf), optarg);
                    } else if (check_params(buf, sizeof(buf), params, optarg) < 0) {
                        fprintf(stderr,
                                "qemu: unknown boot parameter '%s' in '%s'\n",
                                buf, optarg);
                        exit(1);
                    }

                    if (legacy ||
                        get_param_value(buf, sizeof(buf), "order", optarg)) {
                        boot_devices_bitmap = parse_bootdevices(buf);
                        pstrcpy(boot_devices, sizeof(boot_devices), buf);
5195
                    }
J
Jan Kiszka 已提交
5196 5197 5198 5199 5200 5201 5202 5203 5204
                    if (!legacy) {
                        if (get_param_value(buf, sizeof(buf),
                                            "once", optarg)) {
                            boot_devices_bitmap |= parse_bootdevices(buf);
                            standard_boot_devices = qemu_strdup(boot_devices);
                            pstrcpy(boot_devices, sizeof(boot_devices), buf);
                            qemu_register_reset(restore_boot_devices,
                                                standard_boot_devices);
                        }
5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217
                        if (get_param_value(buf, sizeof(buf),
                                            "menu", optarg)) {
                            if (!strcmp(buf, "on")) {
                                boot_menu = 1;
                            } else if (!strcmp(buf, "off")) {
                                boot_menu = 0;
                            } else {
                                fprintf(stderr,
                                        "qemu: invalid option value '%s'\n",
                                        buf);
                                exit(1);
                            }
                        }
J
Jan Kiszka 已提交
5218
                    }
5219 5220
                }
                break;
5221 5222
            case QEMU_OPTION_fda:
            case QEMU_OPTION_fdb:
5223
                drive_add(optarg, FD_ALIAS, popt->index - QEMU_OPTION_fda);
5224
                break;
B
bellard 已提交
5225 5226 5227 5228 5229
#ifdef TARGET_I386
            case QEMU_OPTION_no_fd_bootchk:
                fd_bootchk = 0;
                break;
#endif
M
Mark McLoughlin 已提交
5230 5231 5232 5233 5234
            case QEMU_OPTION_netdev:
                if (net_client_parse(&qemu_netdev_opts, optarg) == -1) {
                    exit(1);
                }
                break;
B
bellard 已提交
5235
            case QEMU_OPTION_net:
5236
                if (net_client_parse(&qemu_net_opts, optarg) == -1) {
5237 5238
                    exit(1);
                }
B
bellard 已提交
5239
                break;
B
bellard 已提交
5240 5241
#ifdef CONFIG_SLIRP
            case QEMU_OPTION_tftp:
5242
                legacy_tftp_prefix = optarg;
B
bellard 已提交
5243
                break;
5244
            case QEMU_OPTION_bootp:
5245
                legacy_bootp_filename = optarg;
5246
                break;
B
bellard 已提交
5247
#ifndef _WIN32
B
bellard 已提交
5248
            case QEMU_OPTION_smb:
M
Markus Armbruster 已提交
5249 5250
                if (net_slirp_smb(optarg) < 0)
                    exit(1);
B
bellard 已提交
5251
                break;
B
bellard 已提交
5252
#endif
B
bellard 已提交
5253
            case QEMU_OPTION_redir:
M
Markus Armbruster 已提交
5254 5255
                if (net_slirp_redir(optarg) < 0)
                    exit(1);
B
bellard 已提交
5256
                break;
B
bellard 已提交
5257
#endif
5258
            case QEMU_OPTION_bt:
5259
                add_device_config(DEV_BT, optarg);
5260
                break;
5261 5262 5263 5264 5265 5266 5267 5268 5269
#ifdef HAS_AUDIO
            case QEMU_OPTION_audio_help:
                AUD_help ();
                exit (0);
                break;
            case QEMU_OPTION_soundhw:
                select_soundhw (optarg);
                break;
#endif
5270
            case QEMU_OPTION_h:
5271
                help(0);
5272
                break;
P
pbrook 已提交
5273 5274 5275 5276
            case QEMU_OPTION_version:
                version();
                exit(0);
                break;
5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290
            case QEMU_OPTION_m: {
                uint64_t value;
                char *ptr;

                value = strtoul(optarg, &ptr, 10);
                switch (*ptr) {
                case 0: case 'M': case 'm':
                    value <<= 20;
                    break;
                case 'G': case 'g':
                    value <<= 30;
                    break;
                default:
                    fprintf(stderr, "qemu: invalid ram size: %s\n", optarg);
5291 5292
                    exit(1);
                }
5293 5294

                /* On 32-bit hosts, QEMU is limited by virtual address space */
5295
                if (value > (2047 << 20) && HOST_LONG_BITS == 32) {
5296 5297 5298
                    fprintf(stderr, "qemu: at most 2047 MB RAM can be simulated\n");
                    exit(1);
                }
A
Anthony Liguori 已提交
5299
                if (value != (uint64_t)(ram_addr_t)value) {
5300 5301 5302 5303
                    fprintf(stderr, "qemu: ram size too large\n");
                    exit(1);
                }
                ram_size = value;
5304
                break;
5305
            }
5306 5307 5308 5309 5310 5311 5312 5313
            case QEMU_OPTION_mempath:
                mem_path = optarg;
                break;
#ifdef MAP_POPULATE
            case QEMU_OPTION_mem_prealloc:
                mem_prealloc = 1;
                break;
#endif
5314 5315 5316
            case QEMU_OPTION_d:
                {
                    int mask;
B
blueswir1 已提交
5317
                    const CPULogItem *item;
5318

5319 5320 5321
                    mask = cpu_str_to_log_mask(optarg);
                    if (!mask) {
                        printf("Log items (comma separated):\n");
5322 5323 5324 5325
                    for(item = cpu_log_items; item->mask != 0; item++) {
                        printf("%-10s %s\n", item->name, item->help);
                    }
                    exit(1);
5326 5327
                    }
                    cpu_set_log(mask);
5328
                }
5329 5330
                break;
            case QEMU_OPTION_s:
5331
                gdbstub_dev = "tcp::" DEFAULT_GDBSTUB_PORT;
5332
                break;
5333 5334
            case QEMU_OPTION_gdb:
                gdbstub_dev = optarg;
5335 5336
                break;
            case QEMU_OPTION_L:
P
Paul Brook 已提交
5337
                data_dir = optarg;
5338
                break;
5339 5340 5341
            case QEMU_OPTION_bios:
                bios_name = optarg;
                break;
5342 5343 5344
            case QEMU_OPTION_singlestep:
                singlestep = 1;
                break;
5345
            case QEMU_OPTION_S:
5346
                autostart = 0;
5347
                break;
5348 5349 5350
	    case QEMU_OPTION_k:
		keyboard_layout = optarg;
		break;
B
bellard 已提交
5351 5352 5353
            case QEMU_OPTION_localtime:
                rtc_utc = 0;
                break;
5354 5355
            case QEMU_OPTION_vga:
                select_vgahw (optarg);
5356
                break;
5357
#if defined(TARGET_PPC) || defined(TARGET_SPARC)
5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377
            case QEMU_OPTION_g:
                {
                    const char *p;
                    int w, h, depth;
                    p = optarg;
                    w = strtol(p, (char **)&p, 10);
                    if (w <= 0) {
                    graphic_error:
                        fprintf(stderr, "qemu: invalid resolution or depth\n");
                        exit(1);
                    }
                    if (*p != 'x')
                        goto graphic_error;
                    p++;
                    h = strtol(p, (char **)&p, 10);
                    if (h <= 0)
                        goto graphic_error;
                    if (*p == 'x') {
                        p++;
                        depth = strtol(p, (char **)&p, 10);
5378
                        if (depth != 8 && depth != 15 && depth != 16 &&
5379 5380 5381 5382 5383 5384 5385
                            depth != 24 && depth != 32)
                            goto graphic_error;
                    } else if (*p == '\0') {
                        depth = graphic_depth;
                    } else {
                        goto graphic_error;
                    }
5386

5387 5388 5389 5390 5391
                    graphic_width = w;
                    graphic_height = h;
                    graphic_depth = depth;
                }
                break;
5392
#endif
T
ths 已提交
5393 5394 5395 5396 5397 5398 5399 5400
            case QEMU_OPTION_echr:
                {
                    char *r;
                    term_escape_char = strtol(optarg, &r, 0);
                    if (r == optarg)
                        printf("Bad argument to echr\n");
                    break;
                }
B
bellard 已提交
5401
            case QEMU_OPTION_monitor:
G
Gerd Hoffmann 已提交
5402 5403 5404 5405 5406
                monitor_parse(optarg, "readline");
                default_monitor = 0;
                break;
            case QEMU_OPTION_qmp:
                monitor_parse(optarg, "control");
G
Gerd Hoffmann 已提交
5407
                default_monitor = 0;
B
bellard 已提交
5408
                break;
G
Gerd Hoffmann 已提交
5409
            case QEMU_OPTION_mon:
5410
                opts = qemu_opts_parse(&qemu_mon_opts, optarg, 1);
G
Gerd Hoffmann 已提交
5411 5412 5413 5414 5415 5416
                if (!opts) {
                    fprintf(stderr, "parse error: %s\n", optarg);
                    exit(1);
                }
                default_monitor = 0;
                break;
5417
            case QEMU_OPTION_chardev:
5418
                opts = qemu_opts_parse(&qemu_chardev_opts, optarg, 1);
5419 5420 5421 5422 5423
                if (!opts) {
                    fprintf(stderr, "parse error: %s\n", optarg);
                    exit(1);
                }
                break;
B
bellard 已提交
5424
            case QEMU_OPTION_serial:
5425 5426
                add_device_config(DEV_SERIAL, optarg);
                default_serial = 0;
5427 5428 5429
                if (strncmp(optarg, "mon:", 4) == 0) {
                    default_monitor = 0;
                }
B
bellard 已提交
5430
                break;
R
Richard W.M. Jones 已提交
5431
            case QEMU_OPTION_watchdog:
M
Markus Armbruster 已提交
5432 5433 5434 5435 5436 5437
                if (watchdog) {
                    fprintf(stderr,
                            "qemu: only one watchdog option may be given\n");
                    return 1;
                }
                watchdog = optarg;
R
Richard W.M. Jones 已提交
5438 5439 5440 5441 5442 5443 5444
                break;
            case QEMU_OPTION_watchdog_action:
                if (select_watchdog_action(optarg) == -1) {
                    fprintf(stderr, "Unknown -watchdog-action parameter\n");
                    exit(1);
                }
                break;
5445
            case QEMU_OPTION_virtiocon:
G
Gerd Hoffmann 已提交
5446 5447
                add_device_config(DEV_VIRTCON, optarg);
                default_virtcon = 0;
5448 5449 5450
                if (strncmp(optarg, "mon:", 4) == 0) {
                    default_monitor = 0;
                }
5451
                break;
5452
            case QEMU_OPTION_parallel:
G
Gerd Hoffmann 已提交
5453 5454
                add_device_config(DEV_PARALLEL, optarg);
                default_parallel = 0;
5455 5456 5457
                if (strncmp(optarg, "mon:", 4) == 0) {
                    default_monitor = 0;
                }
5458
                break;
5459 5460 5461
            case QEMU_OPTION_debugcon:
                add_device_config(DEV_DEBUGCON, optarg);
                break;
B
bellard 已提交
5462 5463 5464 5465 5466 5467
	    case QEMU_OPTION_loadvm:
		loadvm = optarg;
		break;
            case QEMU_OPTION_full_screen:
                full_screen = 1;
                break;
T
ths 已提交
5468
#ifdef CONFIG_SDL
5469 5470 5471
            case QEMU_OPTION_no_frame:
                no_frame = 1;
                break;
T
ths 已提交
5472 5473 5474
            case QEMU_OPTION_alt_grab:
                alt_grab = 1;
                break;
5475 5476 5477
            case QEMU_OPTION_ctrl_grab:
                ctrl_grab = 1;
                break;
T
ths 已提交
5478 5479 5480
            case QEMU_OPTION_no_quit:
                no_quit = 1;
                break;
5481
            case QEMU_OPTION_sdl:
5482
                display_type = DT_SDL;
5483
                break;
T
ths 已提交
5484
#endif
B
bellard 已提交
5485
            case QEMU_OPTION_pidfile:
5486
                pid_file = optarg;
B
bellard 已提交
5487
                break;
5488 5489 5490 5491
#ifdef TARGET_I386
            case QEMU_OPTION_win2k_hack:
                win2k_install_hack = 1;
                break;
5492 5493 5494
            case QEMU_OPTION_rtc_td_hack:
                rtc_td_hack = 1;
                break;
5495 5496 5497 5498 5499 5500
            case QEMU_OPTION_acpitable:
                if(acpi_table_add(optarg) < 0) {
                    fprintf(stderr, "Wrong acpi table provided\n");
                    exit(1);
                }
                break;
5501 5502 5503 5504 5505 5506
            case QEMU_OPTION_smbios:
                if(smbios_entry_add(optarg) < 0) {
                    fprintf(stderr, "Wrong smbios provided\n");
                    exit(1);
                }
                break;
5507
#endif
A
aliguori 已提交
5508 5509 5510 5511
#ifdef CONFIG_KVM
            case QEMU_OPTION_enable_kvm:
                kvm_allowed = 1;
                break;
B
bellard 已提交
5512
#endif
B
bellard 已提交
5513 5514 5515
            case QEMU_OPTION_usb:
                usb_enabled = 1;
                break;
B
bellard 已提交
5516 5517
            case QEMU_OPTION_usbdevice:
                usb_enabled = 1;
5518 5519 5520
                add_device_config(DEV_USB, optarg);
                break;
            case QEMU_OPTION_device:
5521
                if (!qemu_opts_parse(&qemu_device_opts, optarg, 1)) {
G
Gerd Hoffmann 已提交
5522 5523
                    exit(1);
                }
B
bellard 已提交
5524
                break;
B
bellard 已提交
5525
            case QEMU_OPTION_smp:
5526
                smp_parse(optarg);
5527
                if (smp_cpus < 1) {
B
bellard 已提交
5528 5529 5530
                    fprintf(stderr, "Invalid number of CPUs\n");
                    exit(1);
                }
5531 5532 5533 5534 5535 5536 5537 5538 5539
                if (max_cpus < smp_cpus) {
                    fprintf(stderr, "maxcpus must be equal to or greater than "
                            "smp\n");
                    exit(1);
                }
                if (max_cpus > 255) {
                    fprintf(stderr, "Unsupported number of maxcpus\n");
                    exit(1);
                }
B
bellard 已提交
5540
                break;
B
bellard 已提交
5541
	    case QEMU_OPTION_vnc:
5542
                display_type = DT_VNC;
5543
		vnc_display = optarg;
B
bellard 已提交
5544
		break;
5545
#ifdef TARGET_I386
B
bellard 已提交
5546 5547 5548
            case QEMU_OPTION_no_acpi:
                acpi_enabled = 0;
                break;
A
aliguori 已提交
5549 5550 5551
            case QEMU_OPTION_no_hpet:
                no_hpet = 1;
                break;
5552 5553 5554 5555 5556
            case QEMU_OPTION_balloon:
                if (balloon_parse(optarg) < 0) {
                    fprintf(stderr, "Unknown -balloon argument %s\n", optarg);
                    exit(1);
                }
5557
                break;
5558
#endif
B
bellard 已提交
5559 5560 5561
            case QEMU_OPTION_no_reboot:
                no_reboot = 1;
                break;
A
aurel32 已提交
5562 5563 5564
            case QEMU_OPTION_no_shutdown:
                no_shutdown = 1;
                break;
5565 5566 5567
            case QEMU_OPTION_show_cursor:
                cursor_hide = 0;
                break;
5568 5569 5570 5571 5572 5573 5574
            case QEMU_OPTION_uuid:
                if(qemu_uuid_parse(optarg, qemu_uuid) < 0) {
                    fprintf(stderr, "Fail to parse UUID string."
                            " Wrong format.\n");
                    exit(1);
                }
                break;
5575
#ifndef _WIN32
T
ths 已提交
5576 5577 5578
	    case QEMU_OPTION_daemonize:
		daemonize = 1;
		break;
5579
#endif
5580 5581 5582 5583 5584 5585 5586 5587
	    case QEMU_OPTION_option_rom:
		if (nb_option_roms >= MAX_OPTION_ROMS) {
		    fprintf(stderr, "Too many option ROMs\n");
		    exit(1);
		}
		option_rom[nb_option_roms] = optarg;
		nb_option_roms++;
		break;
5588
#if defined(TARGET_ARM) || defined(TARGET_M68K)
5589 5590 5591
            case QEMU_OPTION_semihosting:
                semihosting_enabled = 1;
                break;
5592
#endif
T
ths 已提交
5593
            case QEMU_OPTION_name:
A
Andi Kleen 已提交
5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606
                qemu_name = qemu_strdup(optarg);
		 {
		     char *p = strchr(qemu_name, ',');
		     if (p != NULL) {
		        *p++ = 0;
			if (strncmp(p, "process=", 8)) {
			    fprintf(stderr, "Unknown subargument %s to -name", p);
			    exit(1);
			}
			p += 8;
			set_proc_name(p);
		     }	
		 }	
T
ths 已提交
5607
                break;
5608
#if defined(TARGET_SPARC) || defined(TARGET_PPC)
B
blueswir1 已提交
5609 5610 5611 5612 5613 5614 5615 5616
            case QEMU_OPTION_prom_env:
                if (nb_prom_envs >= MAX_PROM_ENVS) {
                    fprintf(stderr, "Too many prom variables\n");
                    exit(1);
                }
                prom_envs[nb_prom_envs] = optarg;
                nb_prom_envs++;
                break;
5617 5618 5619 5620
#endif
#ifdef TARGET_ARM
            case QEMU_OPTION_old_param:
                old_param = 1;
5621
                break;
B
blueswir1 已提交
5622
#endif
5623 5624 5625
            case QEMU_OPTION_clock:
                configure_alarms(optarg);
                break;
B
bellard 已提交
5626
            case QEMU_OPTION_startdate:
J
Jan Kiszka 已提交
5627 5628 5629
                configure_rtc_date_offset(optarg, 1);
                break;
            case QEMU_OPTION_rtc:
5630
                opts = qemu_opts_parse(&qemu_rtc_opts, optarg, 0);
J
Jan Kiszka 已提交
5631 5632 5633
                if (!opts) {
                    fprintf(stderr, "parse error: %s\n", optarg);
                    exit(1);
B
bellard 已提交
5634
                }
J
Jan Kiszka 已提交
5635
                configure_rtc(opts);
B
bellard 已提交
5636
                break;
5637 5638 5639 5640 5641
            case QEMU_OPTION_tb_size:
                tb_size = strtol(optarg, NULL, 0);
                if (tb_size < 0)
                    tb_size = 0;
                break;
P
pbrook 已提交
5642
            case QEMU_OPTION_icount:
P
Paolo Bonzini 已提交
5643
                icount_option = optarg;
P
pbrook 已提交
5644
                break;
A
aliguori 已提交
5645 5646 5647
            case QEMU_OPTION_incoming:
                incoming = optarg;
                break;
5648 5649 5650
            case QEMU_OPTION_nodefaults:
                default_serial = 0;
                default_parallel = 0;
G
Gerd Hoffmann 已提交
5651
                default_virtcon = 0;
5652 5653
                default_monitor = 0;
                default_vga = 0;
G
Gerd Hoffmann 已提交
5654
                default_net = 0;
G
Gerd Hoffmann 已提交
5655 5656 5657
                default_floppy = 0;
                default_cdrom = 0;
                default_sdcard = 0;
5658
                break;
5659
#ifndef _WIN32
5660 5661 5662 5663 5664 5665
            case QEMU_OPTION_chroot:
                chroot_dir = optarg;
                break;
            case QEMU_OPTION_runas:
                run_as = optarg;
                break;
5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676
#endif
#ifdef CONFIG_XEN
            case QEMU_OPTION_xen_domid:
                xen_domid = atoi(optarg);
                break;
            case QEMU_OPTION_xen_create:
                xen_mode = XEN_CREATE;
                break;
            case QEMU_OPTION_xen_attach:
                xen_mode = XEN_ATTACH;
                break;
5677
#endif
5678 5679 5680 5681 5682 5683 5684 5685
            case QEMU_OPTION_readconfig:
                {
                    FILE *fp;
                    fp = fopen(optarg, "r");
                    if (fp == NULL) {
                        fprintf(stderr, "open %s: %s\n", optarg, strerror(errno));
                        exit(1);
                    }
5686
                    if (qemu_config_parse(fp, optarg) != 0) {
5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707
                        exit(1);
                    }
                    fclose(fp);
                    break;
                }
            case QEMU_OPTION_writeconfig:
                {
                    FILE *fp;
                    if (strcmp(optarg, "-") == 0) {
                        fp = stdout;
                    } else {
                        fp = fopen(optarg, "w");
                        if (fp == NULL) {
                            fprintf(stderr, "open %s: %s\n", optarg, strerror(errno));
                            exit(1);
                        }
                    }
                    qemu_config_write(fp);
                    fclose(fp);
                    break;
                }
5708
            }
5709 5710
        }
    }
5711
    loc_set_none();
5712

P
Paul Brook 已提交
5713 5714 5715 5716 5717 5718 5719 5720 5721 5722
    /* If no data_dir is specified then try to find it relative to the
       executable path.  */
    if (!data_dir) {
        data_dir = find_datadir(argv[0]);
    }
    /* If all else fails use the install patch specified when building.  */
    if (!data_dir) {
        data_dir = CONFIG_QEMU_SHAREDIR;
    }

5723 5724 5725 5726 5727 5728 5729
    /*
     * Default to max_cpus = smp_cpus, in case the user doesn't
     * specify a max_cpus value.
     */
    if (!max_cpus)
        max_cpus = smp_cpus;

B
balrog 已提交
5730
    machine->max_cpus = machine->max_cpus ?: 1; /* Default to UP */
5731 5732 5733 5734 5735 5736 5737
    if (smp_cpus > machine->max_cpus) {
        fprintf(stderr, "Number of SMP cpus requested (%d), exceeds max cpus "
                "supported by machine `%s' (%d)\n", smp_cpus,  machine->name,
                machine->max_cpus);
        exit(1);
    }

5738
    qemu_opts_foreach(&qemu_device_opts, default_driver_check, NULL, 0);
G
Gerd Hoffmann 已提交
5739
    qemu_opts_foreach(&qemu_global_opts, default_driver_check, NULL, 0);
5740

5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752
    if (machine->no_serial) {
        default_serial = 0;
    }
    if (machine->no_parallel) {
        default_parallel = 0;
    }
    if (!machine->use_virtcon) {
        default_virtcon = 0;
    }
    if (machine->no_vga) {
        default_vga = 0;
    }
G
Gerd Hoffmann 已提交
5753 5754 5755 5756 5757 5758 5759 5760 5761
    if (machine->no_floppy) {
        default_floppy = 0;
    }
    if (machine->no_cdrom) {
        default_cdrom = 0;
    }
    if (machine->no_sdcard) {
        default_sdcard = 0;
    }
5762

5763
    if (display_type == DT_NOGRAPHIC) {
G
Gerd Hoffmann 已提交
5764 5765
        if (default_parallel)
            add_device_config(DEV_PARALLEL, "null");
G
Gerd Hoffmann 已提交
5766 5767
        if (default_serial && default_monitor) {
            add_device_config(DEV_SERIAL, "mon:stdio");
5768 5769
        } else if (default_virtcon && default_monitor) {
            add_device_config(DEV_VIRTCON, "mon:stdio");
G
Gerd Hoffmann 已提交
5770 5771 5772
        } else {
            if (default_serial)
                add_device_config(DEV_SERIAL, "stdio");
5773 5774
            if (default_virtcon)
                add_device_config(DEV_VIRTCON, "stdio");
G
Gerd Hoffmann 已提交
5775
            if (default_monitor)
G
Gerd Hoffmann 已提交
5776
                monitor_parse("stdio", "readline");
G
Gerd Hoffmann 已提交
5777
        }
5778 5779 5780
    } else {
        if (default_serial)
            add_device_config(DEV_SERIAL, "vc:80Cx24C");
G
Gerd Hoffmann 已提交
5781 5782
        if (default_parallel)
            add_device_config(DEV_PARALLEL, "vc:80Cx24C");
G
Gerd Hoffmann 已提交
5783
        if (default_monitor)
G
Gerd Hoffmann 已提交
5784
            monitor_parse("vc:80Cx24C", "readline");
5785 5786
        if (default_virtcon)
            add_device_config(DEV_VIRTCON, "vc:80Cx24C");
5787
    }
G
Gerd Hoffmann 已提交
5788 5789
    if (default_vga)
        vga_interface_type = VGA_CIRRUS;
5790

5791 5792 5793
    if (qemu_opts_foreach(&qemu_chardev_opts, chardev_init_func, NULL, 1) != 0)
        exit(1);

T
ths 已提交
5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808
#ifndef _WIN32
    if (daemonize) {
	pid_t pid;

	if (pipe(fds) == -1)
	    exit(1);

	pid = fork();
	if (pid > 0) {
	    uint8_t status;
	    ssize_t len;

	    close(fds[1]);

	again:
5809 5810 5811 5812 5813 5814 5815
            len = read(fds[0], &status, 1);
            if (len == -1 && (errno == EINTR))
                goto again;

            if (len != 1)
                exit(1);
            else if (status == 1) {
5816
                fprintf(stderr, "Could not acquire pidfile: %s\n", strerror(errno));
5817 5818 5819
                exit(1);
            } else
                exit(0);
T
ths 已提交
5820
	} else if (pid < 0)
5821
            exit(1);
T
ths 已提交
5822

K
Kevin Wolf 已提交
5823 5824 5825
	close(fds[0]);
	qemu_set_cloexec(fds[1]);

T
ths 已提交
5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839
	setsid();

	pid = fork();
	if (pid > 0)
	    exit(0);
	else if (pid < 0)
	    exit(1);

	umask(027);

        signal(SIGTSTP, SIG_IGN);
        signal(SIGTTOU, SIG_IGN);
        signal(SIGTTIN, SIG_IGN);
    }
5840
#endif
T
ths 已提交
5841

5842
    if (pid_file && qemu_create_pidfile(pid_file) != 0) {
5843
#ifndef _WIN32
5844 5845
        if (daemonize) {
            uint8_t status = 1;
5846 5847 5848
            if (write(fds[1], &status, 1) != 1) {
                perror("daemonize. Writing to pipe\n");
            }
5849
        } else
5850
#endif
5851
            fprintf(stderr, "Could not acquire pid file: %s\n", strerror(errno));
5852 5853 5854
        exit(1);
    }

M
Marcelo Tosatti 已提交
5855 5856 5857 5858 5859 5860 5861 5862 5863 5864
    if (kvm_enabled()) {
        int ret;

        ret = kvm_init(smp_cpus);
        if (ret < 0) {
            fprintf(stderr, "failed to initialize KVM\n");
            exit(1);
        }
    }

5865 5866 5867 5868
    if (qemu_init_main_loop()) {
        fprintf(stderr, "qemu_init_main_loop failed\n");
        exit(1);
    }
5869
    linux_boot = (kernel_filename != NULL);
B
balrog 已提交
5870

5871 5872 5873 5874 5875 5876 5877 5878 5879 5880
    if (!linux_boot && *kernel_cmdline != '\0') {
        fprintf(stderr, "-append only allowed with -kernel option\n");
        exit(1);
    }

    if (!linux_boot && initrd_filename != NULL) {
        fprintf(stderr, "-initrd only allowed with -kernel option\n");
        exit(1);
    }

5881 5882
#ifndef _WIN32
    /* Win32 doesn't support line-buffering and requires size >= 2 */
B
bellard 已提交
5883
    setvbuf(stdout, NULL, _IOLBF, 0);
5884
#endif
5885

5886 5887 5888 5889
    if (init_timer_alarm() < 0) {
        fprintf(stderr, "could not initialize alarm timer\n");
        exit(1);
    }
P
Paolo Bonzini 已提交
5890
    configure_icount(icount_option);
5891

B
bellard 已提交
5892 5893 5894 5895
#ifdef _WIN32
    socket_init();
#endif

5896 5897
    if (net_init_clients() < 0) {
        exit(1);
B
bellard 已提交
5898
    }
B
bellard 已提交
5899

5900 5901 5902
    net_boot = (boot_devices_bitmap >> ('n' - 'a')) & 0xF;
    net_set_boot_mask(net_boot);

5903
    /* init the bluetooth world */
5904 5905
    if (foreach_device_config(DEV_BT, bt_parse))
        exit(1);
5906

5907
    /* init the memory */
P
pbrook 已提交
5908 5909
    if (ram_size == 0)
        ram_size = DEFAULT_RAM_SIZE * 1024 * 1024;
5910

5911 5912 5913
    /* init the dynamic translator */
    cpu_exec_init_all(tb_size * 1024 * 1024);

5914
    bdrv_init_with_whitelist();
5915

L
lirans@il.ibm.com 已提交
5916 5917
    blk_mig_init();

G
Gerd Hoffmann 已提交
5918
    if (default_cdrom) {
G
Gerd Hoffmann 已提交
5919 5920
        /* we always create the cdrom drive, even if no disk is there */
        drive_add(NULL, CDROM_ALIAS);
G
Gerd Hoffmann 已提交
5921
    }
5922

G
Gerd Hoffmann 已提交
5923
    if (default_floppy) {
G
Gerd Hoffmann 已提交
5924 5925
        /* we always create at least one floppy */
        drive_add(NULL, FD_ALIAS, 0);
G
Gerd Hoffmann 已提交
5926
    }
5927

G
Gerd Hoffmann 已提交
5928
    if (default_sdcard) {
G
Gerd Hoffmann 已提交
5929 5930 5931
        /* we always create one sd slot, even if no card is in it */
        drive_add(NULL, SD_ALIAS);
    }
5932

T
ths 已提交
5933
    /* open the virtual block devices */
G
Gerd Hoffmann 已提交
5934
    if (snapshot)
G
Gerd Hoffmann 已提交
5935 5936
        qemu_opts_foreach(&qemu_drive_opts, drive_enable_snapshot, NULL, 0);
    if (qemu_opts_foreach(&qemu_drive_opts, drive_init_func, machine, 1) != 0)
G
Gerd Hoffmann 已提交
5937
        exit(1);
5938

J
Juan Quintela 已提交
5939
    vmstate_register(0, &vmstate_timers ,&timers_state);
L
lirans@il.ibm.com 已提交
5940 5941
    register_savevm_live("ram", 0, 3, NULL, ram_save_live, NULL, 
                         ram_load, NULL);
5942

5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984
    if (nb_numa_nodes > 0) {
        int i;

        if (nb_numa_nodes > smp_cpus) {
            nb_numa_nodes = smp_cpus;
        }

        /* If no memory size if given for any node, assume the default case
         * and distribute the available memory equally across all nodes
         */
        for (i = 0; i < nb_numa_nodes; i++) {
            if (node_mem[i] != 0)
                break;
        }
        if (i == nb_numa_nodes) {
            uint64_t usedmem = 0;

            /* On Linux, the each node's border has to be 8MB aligned,
             * the final node gets the rest.
             */
            for (i = 0; i < nb_numa_nodes - 1; i++) {
                node_mem[i] = (ram_size / nb_numa_nodes) & ~((1 << 23UL) - 1);
                usedmem += node_mem[i];
            }
            node_mem[i] = ram_size - usedmem;
        }

        for (i = 0; i < nb_numa_nodes; i++) {
            if (node_cpumask[i] != 0)
                break;
        }
        /* assigning the VCPUs round-robin is easier to implement, guest OSes
         * must cope with this anyway, because there are BIOSes out there in
         * real machines which also use this scheme.
         */
        if (i == nb_numa_nodes) {
            for (i = 0; i < smp_cpus; i++) {
                node_cpumask[i % nb_numa_nodes] |= 1 << i;
            }
        }
    }

5985 5986
    if (foreach_device_config(DEV_SERIAL, serial_parse) < 0)
        exit(1);
G
Gerd Hoffmann 已提交
5987 5988
    if (foreach_device_config(DEV_PARALLEL, parallel_parse) < 0)
        exit(1);
G
Gerd Hoffmann 已提交
5989 5990
    if (foreach_device_config(DEV_VIRTCON, virtcon_parse) < 0)
        exit(1);
5991 5992
    if (foreach_device_config(DEV_DEBUGCON, debugcon_parse) < 0)
        exit(1);
5993

P
Paul Brook 已提交
5994 5995
    module_call_init(MODULE_INIT_DEVICE);

5996 5997 5998
    if (qemu_opts_foreach(&qemu_device_opts, device_help_func, NULL, 0) != 0)
        exit(0);

M
Markus Armbruster 已提交
5999 6000 6001 6002 6003 6004
    if (watchdog) {
        i = select_watchdog(watchdog);
        if (i > 0)
            exit (i == 1 ? 1 : 0);
    }

6005
    if (machine->compat_props) {
6006
        qdev_prop_register_global_list(machine->compat_props);
6007
    }
6008 6009
    qemu_add_globals();

P
Paul Brook 已提交
6010
    machine->init(ram_size, boot_devices,
6011 6012
                  kernel_filename, kernel_cmdline, initrd_filename, cpu_model);

6013
    cpu_synchronize_all_post_init();
6014

J
Juan Quintela 已提交
6015 6016 6017 6018 6019
#ifndef _WIN32
    /* must be after terminal init, SDL library changes signal handlers */
    sighandler_setup();
#endif

6020 6021 6022 6023 6024 6025 6026 6027
    for (env = first_cpu; env != NULL; env = env->next_cpu) {
        for (i = 0; i < nb_numa_nodes; i++) {
            if (node_cpumask[i] & (1 << env->cpu_index)) {
                env->numa_node = i;
            }
        }
    }

6028 6029
    current_machine = machine;

6030 6031
    /* init USB devices */
    if (usb_enabled) {
M
Markus Armbruster 已提交
6032 6033
        if (foreach_device_config(DEV_USB, usb_parse) < 0)
            exit(1);
6034 6035
    }

6036
    /* init generic devices */
G
Gerd Hoffmann 已提交
6037
    if (qemu_opts_foreach(&qemu_device_opts, device_init_func, NULL, 1) != 0)
6038 6039
        exit(1);

6040 6041
    net_check_clients();

6042
    /* just use the first displaystate for the moment */
6043
    ds = get_displaystate();
6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058

    if (display_type == DT_DEFAULT) {
#if defined(CONFIG_SDL) || defined(CONFIG_COCOA)
        display_type = DT_SDL;
#else
        display_type = DT_VNC;
        vnc_display = "localhost:0,to=99";
        show_vnc_port = 1;
#endif
    }
        

    switch (display_type) {
    case DT_NOGRAPHIC:
        break;
B
balrog 已提交
6059
#if defined(CONFIG_CURSES)
6060 6061 6062
    case DT_CURSES:
        curses_display_init(ds, full_screen);
        break;
B
balrog 已提交
6063
#endif
6064
#if defined(CONFIG_SDL)
6065 6066 6067
    case DT_SDL:
        sdl_display_init(ds, full_screen, no_frame);
        break;
6068
#elif defined(CONFIG_COCOA)
6069 6070 6071
    case DT_SDL:
        cocoa_display_init(ds, full_screen);
        break;
6072
#endif
6073 6074 6075 6076
    case DT_VNC:
        vnc_display_init(ds);
        if (vnc_display_open(ds, vnc_display) < 0)
            exit(1);
6077

6078 6079
        if (show_vnc_port) {
            printf("VNC server running on `%s'\n", vnc_display_local_addr(ds));
6080
        }
6081 6082 6083
        break;
    default:
        break;
6084
    }
6085
    dpy_resize(ds);
6086

6087 6088 6089 6090 6091
    dcl = ds->listeners;
    while (dcl != NULL) {
        if (dcl->dpy_refresh != NULL) {
            ds->gui_timer = qemu_new_timer(rt_clock, gui_update, ds);
            qemu_mod_timer(ds->gui_timer, qemu_get_clock(rt_clock));
T
ths 已提交
6092
        }
6093
        dcl = dcl->next;
T
ths 已提交
6094
    }
6095

6096
    if (display_type == DT_NOGRAPHIC || display_type == DT_VNC) {
B
blueswir1 已提交
6097 6098 6099 6100
        nographic_timer = qemu_new_timer(rt_clock, nographic_update, NULL);
        qemu_mod_timer(nographic_timer, qemu_get_clock(rt_clock));
    }

6101
    text_consoles_set_display(ds);
6102

6103 6104
    if (qemu_opts_foreach(&qemu_mon_opts, mon_init_func, NULL, 1) != 0)
        exit(1);
B
bellard 已提交
6105

6106 6107 6108 6109
    if (gdbstub_dev && gdbserver_start(gdbstub_dev) < 0) {
        fprintf(stderr, "qemu: could not open gdbserver on device '%s'\n",
                gdbstub_dev);
        exit(1);
6110 6111
    }

6112 6113
    qdev_machine_creation_done();

G
Gerd Hoffmann 已提交
6114 6115 6116 6117
    if (rom_load_all() != 0) {
        fprintf(stderr, "rom loading failed\n");
        exit(1);
    }
6118

6119
    qemu_system_reset();
6120
    if (loadvm) {
6121
        if (load_vmstate(loadvm) < 0) {
6122 6123 6124
            autostart = 0;
        }
    }
B
bellard 已提交
6125

G
Glauber Costa 已提交
6126
    if (incoming) {
A
aliguori 已提交
6127
        qemu_start_incoming_migration(incoming);
6128
    } else if (autostart) {
6129
        vm_start();
6130
    }
6131

B
blueswir1 已提交
6132
#ifndef _WIN32
T
ths 已提交
6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144
    if (daemonize) {
	uint8_t status = 0;
	ssize_t len;

    again1:
	len = write(fds[1], &status, 1);
	if (len == -1 && (errno == EINTR))
	    goto again1;

	if (len != 1)
	    exit(1);

6145 6146 6147 6148
        if (chdir("/")) {
            perror("not able to chdir to /");
            exit(1);
        }
K
Kevin Wolf 已提交
6149
	TFR(fd = qemu_open("/dev/null", O_RDWR));
T
ths 已提交
6150 6151
	if (fd == -1)
	    exit(1);
6152
    }
T
ths 已提交
6153

6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166
    if (run_as) {
        pwd = getpwnam(run_as);
        if (!pwd) {
            fprintf(stderr, "User \"%s\" doesn't exist\n", run_as);
            exit(1);
        }
    }

    if (chroot_dir) {
        if (chroot(chroot_dir) < 0) {
            fprintf(stderr, "chroot failed\n");
            exit(1);
        }
6167 6168 6169 6170
        if (chdir("/")) {
            perror("not able to chdir to /");
            exit(1);
        }
6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191
    }

    if (run_as) {
        if (setgid(pwd->pw_gid) < 0) {
            fprintf(stderr, "Failed to setgid(%d)\n", pwd->pw_gid);
            exit(1);
        }
        if (setuid(pwd->pw_uid) < 0) {
            fprintf(stderr, "Failed to setuid(%d)\n", pwd->pw_uid);
            exit(1);
        }
        if (setuid(0) != -1) {
            fprintf(stderr, "Dropping privileges failed\n");
            exit(1);
        }
    }

    if (daemonize) {
        dup2(fd, 0);
        dup2(fd, 1);
        dup2(fd, 2);
T
ths 已提交
6192

6193
        close(fd);
T
ths 已提交
6194
    }
B
blueswir1 已提交
6195
#endif
T
ths 已提交
6196

6197
    main_loop();
B
bellard 已提交
6198
    quit_timers();
6199
    net_cleanup();
T
ths 已提交
6200

6201 6202
    return 0;
}