arch_init.c 31.9 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
/*
 * QEMU System Emulator
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */
#include <stdint.h>
#include <stdarg.h>
26
#include <stdlib.h>
27
#ifndef _WIN32
28
#include <sys/types.h>
29
30
31
#include <sys/mman.h>
#endif
#include "config.h"
32
#include "monitor/monitor.h"
33
#include "sysemu/sysemu.h"
34
35
#include "qemu/bitops.h"
#include "qemu/bitmap.h"
36
#include "sysemu/arch_init.h"
37
#include "audio/audio.h"
Paolo Bonzini's avatar
Paolo Bonzini committed
38
#include "hw/i386/pc.h"
39
#include "hw/pci/pci.h"
Paolo Bonzini's avatar
Paolo Bonzini committed
40
#include "hw/audio/audio.h"
41
#include "sysemu/kvm.h"
42
#include "migration/migration.h"
Paolo Bonzini's avatar
Paolo Bonzini committed
43
#include "hw/i386/smbios.h"
44
#include "exec/address-spaces.h"
Paolo Bonzini's avatar
Paolo Bonzini committed
45
#include "hw/audio/pcspk.h"
46
#include "migration/page_cache.h"
47
#include "qemu/config-file.h"
48
#include "qmp-commands.h"
49
#include "trace.h"
50
#include "exec/cpu-all.h"
51
#include "hw/acpi/acpi.h"
52

Orit Wasserman's avatar
Orit Wasserman committed
53
54
55
56
57
58
59
60
#ifdef DEBUG_ARCH_INIT
#define DPRINTF(fmt, ...) \
    do { fprintf(stdout, "arch_init: " fmt, ## __VA_ARGS__); } while (0)
#else
#define DPRINTF(fmt, ...) \
    do { } while (0)
#endif

61
62
63
64
65
66
67
#ifdef TARGET_SPARC
int graphic_width = 1024;
int graphic_height = 768;
int graphic_depth = 8;
#else
int graphic_width = 800;
int graphic_height = 600;
68
int graphic_depth = 32;
69
70
71
72
73
74
75
76
77
78
79
80
81
#endif


#if defined(TARGET_ALPHA)
#define QEMU_ARCH QEMU_ARCH_ALPHA
#elif defined(TARGET_ARM)
#define QEMU_ARCH QEMU_ARCH_ARM
#elif defined(TARGET_CRIS)
#define QEMU_ARCH QEMU_ARCH_CRIS
#elif defined(TARGET_I386)
#define QEMU_ARCH QEMU_ARCH_I386
#elif defined(TARGET_M68K)
#define QEMU_ARCH QEMU_ARCH_M68K
Michael Walle's avatar
Michael Walle committed
82
83
#elif defined(TARGET_LM32)
#define QEMU_ARCH QEMU_ARCH_LM32
84
85
86
87
#elif defined(TARGET_MICROBLAZE)
#define QEMU_ARCH QEMU_ARCH_MICROBLAZE
#elif defined(TARGET_MIPS)
#define QEMU_ARCH QEMU_ARCH_MIPS
88
89
#elif defined(TARGET_MOXIE)
#define QEMU_ARCH QEMU_ARCH_MOXIE
90
91
#elif defined(TARGET_OPENRISC)
#define QEMU_ARCH QEMU_ARCH_OPENRISC
92
93
94
95
96
97
98
99
#elif defined(TARGET_PPC)
#define QEMU_ARCH QEMU_ARCH_PPC
#elif defined(TARGET_S390X)
#define QEMU_ARCH QEMU_ARCH_S390X
#elif defined(TARGET_SH4)
#define QEMU_ARCH QEMU_ARCH_SH4
#elif defined(TARGET_SPARC)
#define QEMU_ARCH QEMU_ARCH_SPARC
Max Filippov's avatar
Max Filippov committed
100
101
#elif defined(TARGET_XTENSA)
#define QEMU_ARCH QEMU_ARCH_XTENSA
102
103
#elif defined(TARGET_UNICORE32)
#define QEMU_ARCH QEMU_ARCH_UNICORE32
104
105
106
#endif

const uint32_t arch_type = QEMU_ARCH;
107
108
109
static bool mig_throttle_on;
static int dirty_rate_high_cnt;
static void check_guest_throttling(void);
110
111
112
113

/***********************************************************/
/* ram save/restore */

114
115
116
117
118
119
#define RAM_SAVE_FLAG_FULL     0x01 /* Obsolete, not used anymore */
#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
#define RAM_SAVE_FLAG_CONTINUE 0x20
120
#define RAM_SAVE_FLAG_XBZRLE   0x40
121

122

123
124
static struct defconfig_file {
    const char *filename;
125
126
    /* Indicates it is an user config file (disabled by -no-user-config) */
    bool userconfig;
127
} default_config_files[] = {
128
    { CONFIG_QEMU_CONFDIR "/qemu.conf",                   true },
129
    { CONFIG_QEMU_CONFDIR "/target-" TARGET_NAME ".conf", true },
130
131
132
133
    { NULL }, /* end of list */
};


134
int qemu_read_default_config_files(bool userconfig)
135
136
{
    int ret;
137
    struct defconfig_file *f;
138

139
    for (f = default_config_files; f->filename; f++) {
140
141
142
        if (!userconfig && f->userconfig) {
            continue;
        }
143
144
145
146
        ret = qemu_read_config_file(f->filename);
        if (ret < 0 && ret != -ENOENT) {
            return ret;
        }
147
    }
Laszlo Ersek's avatar
Laszlo Ersek committed
148

149
150
151
    return 0;
}

152
static inline bool is_zero_page(uint8_t *p)
153
{
154
155
    return buffer_find_nonzero_offset(p, TARGET_PAGE_SIZE) ==
        TARGET_PAGE_SIZE;
156
157
}

158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
/* struct contains XBZRLE cache and a static page
   used by the compression */
static struct {
    /* buffer used for XBZRLE encoding */
    uint8_t *encoded_buf;
    /* buffer for storing page content */
    uint8_t *current_buf;
    /* buffer used for XBZRLE decoding */
    uint8_t *decoded_buf;
    /* Cache for XBZRLE */
    PageCache *cache;
} XBZRLE = {
    .encoded_buf = NULL,
    .current_buf = NULL,
    .decoded_buf = NULL,
    .cache = NULL,
};

176
177
178
179
180
181
182
183
184
185

int64_t xbzrle_cache_resize(int64_t new_size)
{
    if (XBZRLE.cache != NULL) {
        return cache_resize(XBZRLE.cache, new_size / TARGET_PAGE_SIZE) *
            TARGET_PAGE_SIZE;
    }
    return pow2floor(new_size);
}

186
187
188
/* accounting for migration statistics */
typedef struct AccountingInfo {
    uint64_t dup_pages;
189
    uint64_t skipped_pages;
190
191
    uint64_t norm_pages;
    uint64_t iterations;
Orit Wasserman's avatar
Orit Wasserman committed
192
193
194
195
    uint64_t xbzrle_bytes;
    uint64_t xbzrle_pages;
    uint64_t xbzrle_cache_miss;
    uint64_t xbzrle_overflows;
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
} AccountingInfo;

static AccountingInfo acct_info;

static void acct_clear(void)
{
    memset(&acct_info, 0, sizeof(acct_info));
}

uint64_t dup_mig_bytes_transferred(void)
{
    return acct_info.dup_pages * TARGET_PAGE_SIZE;
}

uint64_t dup_mig_pages_transferred(void)
{
    return acct_info.dup_pages;
}

215
216
217
218
219
220
221
222
223
224
uint64_t skipped_mig_bytes_transferred(void)
{
    return acct_info.skipped_pages * TARGET_PAGE_SIZE;
}

uint64_t skipped_mig_pages_transferred(void)
{
    return acct_info.skipped_pages;
}

225
226
227
228
229
230
231
232
233
234
uint64_t norm_mig_bytes_transferred(void)
{
    return acct_info.norm_pages * TARGET_PAGE_SIZE;
}

uint64_t norm_mig_pages_transferred(void)
{
    return acct_info.norm_pages;
}

Orit Wasserman's avatar
Orit Wasserman committed
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
uint64_t xbzrle_mig_bytes_transferred(void)
{
    return acct_info.xbzrle_bytes;
}

uint64_t xbzrle_mig_pages_transferred(void)
{
    return acct_info.xbzrle_pages;
}

uint64_t xbzrle_mig_pages_cache_miss(void)
{
    return acct_info.xbzrle_cache_miss;
}

uint64_t xbzrle_mig_pages_overflow(void)
{
    return acct_info.xbzrle_overflows;
}

255
256
static size_t save_block_hdr(QEMUFile *f, RAMBlock *block, ram_addr_t offset,
                             int cont, int flag)
Orit Wasserman's avatar
Orit Wasserman committed
257
{
258
259
260
261
    size_t size;

    qemu_put_be64(f, offset | cont | flag);
    size = 8;
Orit Wasserman's avatar
Orit Wasserman committed
262

263
264
265
266
267
268
269
    if (!cont) {
        qemu_put_byte(f, strlen(block->idstr));
        qemu_put_buffer(f, (uint8_t *)block->idstr,
                        strlen(block->idstr));
        size += 1 + strlen(block->idstr);
    }
    return size;
Orit Wasserman's avatar
Orit Wasserman committed
270
271
}

272
273
274
275
#define ENCODING_FLAG_XBZRLE 0x1

static int save_xbzrle_page(QEMUFile *f, uint8_t *current_data,
                            ram_addr_t current_addr, RAMBlock *block,
276
                            ram_addr_t offset, int cont, bool last_stage)
277
278
279
280
281
{
    int encoded_len = 0, bytes_sent = -1;
    uint8_t *prev_cached_page;

    if (!cache_is_cached(XBZRLE.cache, current_addr)) {
282
        if (!last_stage) {
283
            cache_insert(XBZRLE.cache, current_addr, current_data);
284
        }
Orit Wasserman's avatar
Orit Wasserman committed
285
        acct_info.xbzrle_cache_miss++;
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
        return -1;
    }

    prev_cached_page = get_cached_data(XBZRLE.cache, current_addr);

    /* save current buffer into memory */
    memcpy(XBZRLE.current_buf, current_data, TARGET_PAGE_SIZE);

    /* XBZRLE encoding (if there is no overflow) */
    encoded_len = xbzrle_encode_buffer(prev_cached_page, XBZRLE.current_buf,
                                       TARGET_PAGE_SIZE, XBZRLE.encoded_buf,
                                       TARGET_PAGE_SIZE);
    if (encoded_len == 0) {
        DPRINTF("Skipping unmodified page\n");
        return 0;
    } else if (encoded_len == -1) {
        DPRINTF("Overflow\n");
Orit Wasserman's avatar
Orit Wasserman committed
303
        acct_info.xbzrle_overflows++;
304
305
306
307
308
309
        /* update data in the cache */
        memcpy(prev_cached_page, current_data, TARGET_PAGE_SIZE);
        return -1;
    }

    /* we need to update the data in the cache, in order to get the same data */
310
311
312
    if (!last_stage) {
        memcpy(prev_cached_page, XBZRLE.current_buf, TARGET_PAGE_SIZE);
    }
313
314

    /* Send XBZRLE based compressed page */
315
    bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_XBZRLE);
316
317
318
    qemu_put_byte(f, ENCODING_FLAG_XBZRLE);
    qemu_put_be16(f, encoded_len);
    qemu_put_buffer(f, XBZRLE.encoded_buf, encoded_len);
319
    bytes_sent += encoded_len + 1 + 2;
Orit Wasserman's avatar
Orit Wasserman committed
320
321
    acct_info.xbzrle_pages++;
    acct_info.xbzrle_bytes += bytes_sent;
322
323
324
325

    return bytes_sent;
}

326
327
328
329

/* This is the last block that we have visited serching for dirty pages
 */
static RAMBlock *last_seen_block;
Juan Quintela's avatar
Juan Quintela committed
330
331
/* This is the last block from where we have sent data */
static RAMBlock *last_sent_block;
332
static ram_addr_t last_offset;
Juan Quintela's avatar
Juan Quintela committed
333
334
static unsigned long *migration_bitmap;
static uint64_t migration_dirty_pages;
335
static uint32_t last_version;
336
static bool ram_bulk_stage;
337

338
339
340
static inline
ram_addr_t migration_bitmap_find_and_reset_dirty(MemoryRegion *mr,
                                                 ram_addr_t start)
341
{
342
343
344
    unsigned long base = mr->ram_addr >> TARGET_PAGE_BITS;
    unsigned long nr = base + (start >> TARGET_PAGE_BITS);
    unsigned long size = base + (int128_get64(mr->size) >> TARGET_PAGE_BITS);
Juan Quintela's avatar
Juan Quintela committed
345

346
347
348
349
350
351
352
    unsigned long next;

    if (ram_bulk_stage && nr > base) {
        next = nr + 1;
    } else {
        next = find_next_bit(migration_bitmap, size, nr);
    }
353

354
355
    if (next < size) {
        clear_bit(next, migration_bitmap);
Juan Quintela's avatar
Juan Quintela committed
356
        migration_dirty_pages--;
357
    }
358
    return (next - base) << TARGET_PAGE_BITS;
359
360
}

Juan Quintela's avatar
Juan Quintela committed
361
362
static inline bool migration_bitmap_set_dirty(MemoryRegion *mr,
                                              ram_addr_t offset)
363
{
Juan Quintela's avatar
Juan Quintela committed
364
365
    bool ret;
    int nr = (mr->ram_addr + offset) >> TARGET_PAGE_BITS;
366

Juan Quintela's avatar
Juan Quintela committed
367
368
369
370
    ret = test_and_set_bit(nr, migration_bitmap);

    if (!ret) {
        migration_dirty_pages++;
371
    }
Juan Quintela's avatar
Juan Quintela committed
372
    return ret;
373
374
}

375
376
/* Needs iothread lock! */

377
378
static void migration_bitmap_sync(void)
{
Juan Quintela's avatar
Juan Quintela committed
379
380
381
    RAMBlock *block;
    ram_addr_t addr;
    uint64_t num_dirty_pages_init = migration_dirty_pages;
382
383
    MigrationState *s = migrate_get_current();
    static int64_t start_time;
384
    static int64_t bytes_xfer_prev;
385
386
    static int64_t num_dirty_pages_period;
    int64_t end_time;
387
388
389
390
391
    int64_t bytes_xfer_now;

    if (!bytes_xfer_prev) {
        bytes_xfer_prev = ram_bytes_transferred();
    }
392
393
394
395

    if (!start_time) {
        start_time = qemu_get_clock_ms(rt_clock);
    }
396
397

    trace_migration_bitmap_sync_start();
398
    address_space_sync_dirty_bitmap(&address_space_memory);
Juan Quintela's avatar
Juan Quintela committed
399

400
    QTAILQ_FOREACH(block, &ram_list.blocks, next) {
Juan Quintela's avatar
Juan Quintela committed
401
        for (addr = 0; addr < block->length; addr += TARGET_PAGE_SIZE) {
402
403
404
            if (memory_region_test_and_clear_dirty(block->mr,
                                                   addr, TARGET_PAGE_SIZE,
                                                   DIRTY_MEMORY_MIGRATION)) {
Juan Quintela's avatar
Juan Quintela committed
405
406
407
408
409
                migration_bitmap_set_dirty(block->mr, addr);
            }
        }
    }
    trace_migration_bitmap_sync_end(migration_dirty_pages
410
                                    - num_dirty_pages_init);
411
412
413
414
415
    num_dirty_pages_period += migration_dirty_pages - num_dirty_pages_init;
    end_time = qemu_get_clock_ms(rt_clock);

    /* more than 1 second = 1000 millisecons */
    if (end_time > start_time + 1000) {
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
        if (migrate_auto_converge()) {
            /* The following detection logic can be refined later. For now:
               Check to see if the dirtied bytes is 50% more than the approx.
               amount of bytes that just got transferred since the last time we
               were in this routine. If that happens >N times (for now N==4)
               we turn on the throttle down logic */
            bytes_xfer_now = ram_bytes_transferred();
            if (s->dirty_pages_rate &&
               (num_dirty_pages_period * TARGET_PAGE_SIZE >
                   (bytes_xfer_now - bytes_xfer_prev)/2) &&
               (dirty_rate_high_cnt++ > 4)) {
                    trace_migration_throttle();
                    mig_throttle_on = true;
                    dirty_rate_high_cnt = 0;
             }
             bytes_xfer_prev = bytes_xfer_now;
        } else {
             mig_throttle_on = false;
        }
435
436
        s->dirty_pages_rate = num_dirty_pages_period * 1000
            / (end_time - start_time);
437
        s->dirty_bytes_rate = s->dirty_pages_rate * TARGET_PAGE_SIZE;
438
439
440
        start_time = end_time;
        num_dirty_pages_period = 0;
    }
441
442
}

443
444
445
/*
 * ram_save_block: Writes a page of memory to the stream f
 *
446
447
 * Returns:  The number of bytes written.
 *           0 means no dirty pages
448
449
 */

450
static int ram_save_block(QEMUFile *f, bool last_stage)
451
{
452
    RAMBlock *block = last_seen_block;
453
    ram_addr_t offset = last_offset;
454
    bool complete_round = false;
455
    int bytes_sent = 0;
Avi Kivity's avatar
Avi Kivity committed
456
    MemoryRegion *mr;
457
    ram_addr_t current_addr;
458

459
    if (!block)
460
        block = QTAILQ_FIRST(&ram_list.blocks);
461

462
    while (true) {
Avi Kivity's avatar
Avi Kivity committed
463
        mr = block->mr;
464
465
466
467
468
469
470
471
472
473
474
        offset = migration_bitmap_find_and_reset_dirty(mr, offset);
        if (complete_round && block == last_seen_block &&
            offset >= last_offset) {
            break;
        }
        if (offset >= block->length) {
            offset = 0;
            block = QTAILQ_NEXT(block, next);
            if (!block) {
                block = QTAILQ_FIRST(&ram_list.blocks);
                complete_round = true;
475
                ram_bulk_stage = false;
476
477
            }
        } else {
478
            uint8_t *p;
Juan Quintela's avatar
Juan Quintela committed
479
            int cont = (block == last_sent_block) ?
480
                RAM_SAVE_FLAG_CONTINUE : 0;
481

Avi Kivity's avatar
Avi Kivity committed
482
            p = memory_region_get_ram_ptr(mr) + offset;
483

484
485
            /* In doubt sent page as normal */
            bytes_sent = -1;
486
            if (is_zero_page(p)) {
487
                acct_info.dup_pages++;
488
489
490
491
                bytes_sent = save_block_hdr(f, block, offset, cont,
                                            RAM_SAVE_FLAG_COMPRESS);
                qemu_put_byte(f, 0);
                bytes_sent++;
492
            } else if (!ram_bulk_stage && migrate_use_xbzrle()) {
493
494
                current_addr = block->offset + offset;
                bytes_sent = save_xbzrle_page(f, p, current_addr, block,
495
496
497
498
                                              offset, cont, last_stage);
                if (!last_stage) {
                    p = get_cached_data(XBZRLE.cache, current_addr);
                }
499
500
            }

501
            /* XBZRLE overflow or normal page */
502
            if (bytes_sent == -1) {
503
                bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
504
                qemu_put_buffer_async(f, p, TARGET_PAGE_SIZE);
505
                bytes_sent += TARGET_PAGE_SIZE;
506
                acct_info.norm_pages++;
507
508
            }

509
            /* if page is unmodified, continue to the next */
510
            if (bytes_sent > 0) {
Juan Quintela's avatar
Juan Quintela committed
511
                last_sent_block = block;
512
513
                break;
            }
514
        }
515
    }
516
    last_seen_block = block;
517
    last_offset = offset;
518

519
    return bytes_sent;
520
521
522
523
}

static uint64_t bytes_transferred;

524
525
526
527
528
529
530
531
532
533
534
535
void acct_update_position(QEMUFile *f, size_t size, bool zero)
{
    uint64_t pages = size / TARGET_PAGE_SIZE;
    if (zero) {
        acct_info.dup_pages += pages;
    } else {
        acct_info.norm_pages += pages;
        bytes_transferred += size;
        qemu_update_position(f, size);
    }
}

536
537
static ram_addr_t ram_save_remaining(void)
{
Juan Quintela's avatar
Juan Quintela committed
538
    return migration_dirty_pages;
539
540
541
542
543
544
545
546
547
548
549
550
551
552
}

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)
{
553
554
555
    RAMBlock *block;
    uint64_t total = 0;

556
    QTAILQ_FOREACH(block, &ram_list.blocks, next)
557
558
559
        total += block->length;

    return total;
560
561
}

Orit Wasserman's avatar
Orit Wasserman committed
562
563
static void migration_end(void)
{
564
565
566
567
568
    if (migration_bitmap) {
        memory_global_dirty_log_stop();
        g_free(migration_bitmap);
        migration_bitmap = NULL;
    }
569

570
    if (XBZRLE.cache) {
571
572
573
574
575
576
577
        cache_fini(XBZRLE.cache);
        g_free(XBZRLE.cache);
        g_free(XBZRLE.encoded_buf);
        g_free(XBZRLE.current_buf);
        g_free(XBZRLE.decoded_buf);
        XBZRLE.cache = NULL;
    }
Orit Wasserman's avatar
Orit Wasserman committed
578
579
}

580
581
582
583
584
static void ram_migration_cancel(void *opaque)
{
    migration_end();
}

585
586
static void reset_ram_globals(void)
{
587
    last_seen_block = NULL;
Juan Quintela's avatar
Juan Quintela committed
588
    last_sent_block = NULL;
589
    last_offset = 0;
590
    last_version = ram_list.version;
591
    ram_bulk_stage = true;
592
593
}

594
595
#define MAX_WAIT 50 /* ms, half buffered_file limit */

596
static int ram_save_setup(QEMUFile *f, void *opaque)
597
{
598
    RAMBlock *block;
Juan Quintela's avatar
Juan Quintela committed
599
600
601
    int64_t ram_pages = last_ram_offset() >> TARGET_PAGE_BITS;

    migration_bitmap = bitmap_new(ram_pages);
602
    bitmap_set(migration_bitmap, 0, ram_pages);
Juan Quintela's avatar
Juan Quintela committed
603
    migration_dirty_pages = ram_pages;
604
605
    mig_throttle_on = false;
    dirty_rate_high_cnt = 0;
606

607
608
609
610
611
612
613
614
615
616
    if (migrate_use_xbzrle()) {
        XBZRLE.cache = cache_init(migrate_xbzrle_cache_size() /
                                  TARGET_PAGE_SIZE,
                                  TARGET_PAGE_SIZE);
        if (!XBZRLE.cache) {
            DPRINTF("Error creating cache\n");
            return -1;
        }
        XBZRLE.encoded_buf = g_malloc0(TARGET_PAGE_SIZE);
        XBZRLE.current_buf = g_malloc(TARGET_PAGE_SIZE);
617
        acct_clear();
618
619
    }

620
621
622
623
624
    qemu_mutex_lock_iothread();
    qemu_mutex_lock_ramlist();
    bytes_transferred = 0;
    reset_ram_globals();

625
    memory_global_dirty_log_start();
Juan Quintela's avatar
Juan Quintela committed
626
    migration_bitmap_sync();
627
    qemu_mutex_unlock_iothread();
628

629
    qemu_put_be64(f, ram_bytes_total() | RAM_SAVE_FLAG_MEM_SIZE);
630

631
    QTAILQ_FOREACH(block, &ram_list.blocks, next) {
632
633
634
        qemu_put_byte(f, strlen(block->idstr));
        qemu_put_buffer(f, (uint8_t *)block->idstr, strlen(block->idstr));
        qemu_put_be64(f, block->length);
635
636
    }

637
    qemu_mutex_unlock_ramlist();
638
639
640
641
642
    qemu_put_be64(f, RAM_SAVE_FLAG_EOS);

    return 0;
}

643
static int ram_save_iterate(QEMUFile *f, void *opaque)
644
645
646
{
    int ret;
    int i;
647
    int64_t t0;
648
    int total_sent = 0;
649

650
651
    qemu_mutex_lock_ramlist();

652
653
654
655
    if (ram_list.version != last_version) {
        reset_ram_globals();
    }

656
    t0 = qemu_get_clock_ns(rt_clock);
657
    i = 0;
658
    while ((ret = qemu_file_rate_limit(f)) == 0) {
659
        int bytes_sent;
660

661
        bytes_sent = ram_save_block(f, false);
662
        /* no more blocks to sent */
663
        if (bytes_sent == 0) {
664
665
            break;
        }
666
        total_sent += bytes_sent;
667
        acct_info.iterations++;
668
        check_guest_throttling();
669
670
671
672
673
674
        /* we want to check in the 1st loop, just in case it was the 1st time
           and we had to sync the dirty bitmap.
           qemu_get_clock_ns() is a bit expensive, so we only check each some
           iterations
        */
        if ((i & 63) == 0) {
675
            uint64_t t1 = (qemu_get_clock_ns(rt_clock) - t0) / 1000000;
676
            if (t1 > MAX_WAIT) {
677
                DPRINTF("big wait: %" PRIu64 " milliseconds, %d iterations\n",
678
679
680
681
682
                        t1, i);
                break;
            }
        }
        i++;
683
684
    }

685
686
    qemu_mutex_unlock_ramlist();

687
    if (ret < 0) {
688
        bytes_transferred += total_sent;
689
690
691
        return ret;
    }

692
    qemu_put_be64(f, RAM_SAVE_FLAG_EOS);
693
694
    total_sent += 8;
    bytes_transferred += total_sent;
695

696
    return total_sent;
697
698
699
700
}

static int ram_save_complete(QEMUFile *f, void *opaque)
{
701
    qemu_mutex_lock_ramlist();
702
    migration_bitmap_sync();
703

704
    /* try transferring iterative blocks of memory */
Orit Wasserman's avatar
Orit Wasserman committed
705

706
    /* flush all remaining blocks regardless of rate limiting */
707
    while (true) {
708
709
        int bytes_sent;

710
        bytes_sent = ram_save_block(f, true);
711
        /* no more blocks to sent */
712
        if (bytes_sent == 0) {
713
            break;
714
        }
715
        bytes_transferred += bytes_sent;
716
    }
717
    migration_end();
718

719
    qemu_mutex_unlock_ramlist();
720
721
    qemu_put_be64(f, RAM_SAVE_FLAG_EOS);

722
    return 0;
723
724
}

725
726
727
728
729
730
731
static uint64_t ram_save_pending(QEMUFile *f, void *opaque, uint64_t max_size)
{
    uint64_t remaining_size;

    remaining_size = ram_save_remaining() * TARGET_PAGE_SIZE;

    if (remaining_size < max_size) {
732
        qemu_mutex_lock_iothread();
733
        migration_bitmap_sync();
734
        qemu_mutex_unlock_iothread();
735
736
737
738
739
        remaining_size = ram_save_remaining() * TARGET_PAGE_SIZE;
    }
    return remaining_size;
}

740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
static int load_xbzrle(QEMUFile *f, ram_addr_t addr, void *host)
{
    int ret, rc = 0;
    unsigned int xh_len;
    int xh_flags;

    if (!XBZRLE.decoded_buf) {
        XBZRLE.decoded_buf = g_malloc(TARGET_PAGE_SIZE);
    }

    /* extract RLE header */
    xh_flags = qemu_get_byte(f);
    xh_len = qemu_get_be16(f);

    if (xh_flags != ENCODING_FLAG_XBZRLE) {
        fprintf(stderr, "Failed to load XBZRLE page - wrong compression!\n");
        return -1;
    }

    if (xh_len > TARGET_PAGE_SIZE) {
        fprintf(stderr, "Failed to load XBZRLE page - len overflow!\n");
        return -1;
    }
    /* load data and decode */
    qemu_get_buffer(f, XBZRLE.decoded_buf, xh_len);

    /* decode RLE */
    ret = xbzrle_decode_buffer(XBZRLE.decoded_buf, xh_len, host,
                               TARGET_PAGE_SIZE);
    if (ret == -1) {
        fprintf(stderr, "Failed to load XBZRLE page - decode error!\n");
        rc = -1;
    } else  if (ret > TARGET_PAGE_SIZE) {
        fprintf(stderr, "Failed to load XBZRLE page - size %d exceeds %d!\n",
                ret, TARGET_PAGE_SIZE);
        abort();
    }

    return rc;
}

781
782
783
784
785
786
787
788
789
790
791
792
793
794
static inline void *host_from_stream_offset(QEMUFile *f,
                                            ram_addr_t offset,
                                            int flags)
{
    static RAMBlock *block = NULL;
    char id[256];
    uint8_t len;

    if (flags & RAM_SAVE_FLAG_CONTINUE) {
        if (!block) {
            fprintf(stderr, "Ack, bad migration stream!\n");
            return NULL;
        }

795
        return memory_region_get_ram_ptr(block->mr) + offset;
796
797
798
799
800
801
    }

    len = qemu_get_byte(f);
    qemu_get_buffer(f, (uint8_t *)id, len);
    id[len] = 0;

802
    QTAILQ_FOREACH(block, &ram_list.blocks, next) {
803
        if (!strncmp(id, block->idstr, sizeof(id)))
804
            return memory_region_get_ram_ptr(block->mr) + offset;
805
806
807
808
809
810
    }

    fprintf(stderr, "Can't find block %s!\n", id);
    return NULL;
}

811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
/*
 * If a page (or a whole RDMA chunk) has been
 * determined to be zero, then zap it.
 */
void ram_handle_compressed(void *host, uint8_t ch, uint64_t size)
{
    if (ch != 0 || !is_zero_page(host)) {
        memset(host, ch, size);
#ifndef _WIN32
        if (ch == 0 &&
            (!kvm_enabled() || kvm_has_sync_mmu()) &&
            getpagesize() <= TARGET_PAGE_SIZE) {
            qemu_madvise(host, TARGET_PAGE_SIZE, QEMU_MADV_DONTNEED);
        }
#endif
    }
}

829
static int ram_load(QEMUFile *f, void *opaque, int version_id)
830
831
{
    ram_addr_t addr;
Orit Wasserman's avatar
Orit Wasserman committed
832
    int flags, ret = 0;
833
    int error;
Orit Wasserman's avatar
Orit Wasserman committed
834
835
836
    static uint64_t seq_iter;

    seq_iter++;
837

838
    if (version_id < 4 || version_id > 4) {
839
840
841
842
843
844
845
846
847
848
        return -EINVAL;
    }

    do {
        addr = qemu_get_be64(f);

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

        if (flags & RAM_SAVE_FLAG_MEM_SIZE) {
849
            if (version_id == 4) {
850
851
852
853
854
855
856
857
858
859
860
861
862
863
                /* Synchronize RAM block list */
                char id[256];
                ram_addr_t length;
                ram_addr_t total_ram_bytes = addr;

                while (total_ram_bytes) {
                    RAMBlock *block;
                    uint8_t len;

                    len = qemu_get_byte(f);
                    qemu_get_buffer(f, (uint8_t *)id, len);
                    id[len] = 0;
                    length = qemu_get_be64(f);

864
                    QTAILQ_FOREACH(block, &ram_list.blocks, next) {
865
                        if (!strncmp(id, block->idstr, sizeof(id))) {
Orit Wasserman's avatar
Orit Wasserman committed
866
                            if (block->length != length) {
867
868
869
                                fprintf(stderr,
                                        "Length mismatch: %s: " RAM_ADDR_FMT
                                        " in != " RAM_ADDR_FMT "\n", id, length,
870
                                        block->length);
Orit Wasserman's avatar
Orit Wasserman committed
871
872
873
                                ret =  -EINVAL;
                                goto done;
                            }
874
875
876
877
878
                            break;
                        }
                    }

                    if (!block) {
879
880
                        fprintf(stderr, "Unknown ramblock \"%s\", cannot "
                                "accept migration\n", id);
Orit Wasserman's avatar
Orit Wasserman committed
881
882
                        ret = -EINVAL;
                        goto done;
883
884
885
886
                    }

                    total_ram_bytes -= length;
                }
887
888
889
890
            }
        }

        if (flags & RAM_SAVE_FLAG_COMPRESS) {
891
892
893
            void *host;
            uint8_t ch;

894
            host = host_from_stream_offset(f, addr, flags);
895
896
897
            if (!host) {
                return -EINVAL;
            }
898
899

            ch = qemu_get_byte(f);
900
            ram_handle_compressed(host, ch, TARGET_PAGE_SIZE);
901
        } else if (flags & RAM_SAVE_FLAG_PAGE) {
902
903
            void *host;

904
            host = host_from_stream_offset(f, addr, flags);
905
906
907
            if (!host) {
                return -EINVAL;
            }
908
909

            qemu_get_buffer(f, host, TARGET_PAGE_SIZE);
910
911
912
913
914
915
916
917
918
919
        } else if (flags & RAM_SAVE_FLAG_XBZRLE) {
            void *host = host_from_stream_offset(f, addr, flags);
            if (!host) {
                return -EINVAL;
            }

            if (load_xbzrle(f, addr, host) < 0) {
                ret = -EINVAL;
                goto done;
            }
920
        }
921
922
        error = qemu_file_get_error(f);
        if (error) {
Orit Wasserman's avatar
Orit Wasserman committed
923
924
            ret = error;
            goto done;
925
926
927
        }
    } while (!(flags & RAM_SAVE_FLAG_EOS));

Orit Wasserman's avatar
Orit Wasserman committed
928
done:
929
930
    DPRINTF("Completed load of VM with exit code %d seq iteration "
            "%" PRIu64 "\n", ret, seq_iter);
Orit Wasserman's avatar
Orit Wasserman committed
931
    return ret;
932
933
}

934
SaveVMHandlers savevm_ram_handlers = {
935
    .save_live_setup = ram_save_setup,
936
937
    .save_live_iterate = ram_save_iterate,
    .save_live_complete = ram_save_complete,
938
    .save_live_pending = ram_save_pending,
939
    .load_state = ram_load,
940
    .cancel = ram_migration_cancel,
941
942
};

943
944
945
946
947
948
struct soundhw {
    const char *name;
    const char *descr;
    int enabled;
    int isa;
    union {
949
        int (*init_isa) (ISABus *bus);
950
951
952
953
        int (*init_pci) (PCIBus *bus);
    } init;
};

954
955
static struct soundhw soundhw[9];
static int soundhw_count;
956

957
958
959
960
961
962
963
964
965
966
void isa_register_soundhw(const char *name, const char *descr,
                          int (*init_isa)(ISABus *bus))
{
    assert(soundhw_count < ARRAY_SIZE(soundhw) - 1);
    soundhw[soundhw_count].name = name;
    soundhw[soundhw_count].descr = descr;
    soundhw[soundhw_count].isa = 1;
    soundhw[soundhw_count].init.init_isa = init_isa;
    soundhw_count++;
}
967

968
969
970
971
972
973
974
975
976
977
void pci_register_soundhw(const char *name, const char *descr,
                          int (*init_pci)(PCIBus *bus))
{
    assert(soundhw_count < ARRAY_SIZE(soundhw) - 1);
    soundhw[soundhw_count].name = name;
    soundhw[soundhw_count].descr = descr;
    soundhw[soundhw_count].isa = 0;
    soundhw[soundhw_count].init.init_pci = init_pci;
    soundhw_count++;
}
978
979
980
981
982

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

983
    if (is_help_option(optarg)) {
984
985
    show_valid_cards:

986
987
988
989
990
991
992
993
994
        if (soundhw_count) {
             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");
        } else {
             printf("Machine has no user-selectable audio hardware "
                    "(it may or may not have always-present audio hardware).\n");
995
        }
996
        exit(!is_help_option(optarg));
997
998
999
1000
1001
1002
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
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
    }
    else {
        size_t l;
        const char *p;
        char *e;
        int bad_card = 0;

        if (!strcmp(optarg, "all")) {
            for (c = soundhw; c->name; ++c) {
                c->enabled = 1;
            }
            return;
        }

        p = optarg;
        while (*p) {
            e = strchr(p, ',');
            l = !e ? strlen(p) : (size_t) (e - p);

            for (c = soundhw; c->name; ++c) {
                if (!strncmp(c->name, p, l) && !c->name[l]) {
                    c->enabled = 1;
                    break;
                }
            }

            if (!c->name) {
                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;
        }
    }
}
1042

1043
void audio_init(void)
1044
1045
{
    struct soundhw *c;
1046
1047
    ISABus *isa_bus = (ISABus *) object_resolve_path_type("", TYPE_ISA_BUS, NULL);
    PCIBus *pci_bus = (PCIBus *) object_resolve_path_type("", TYPE_PCI_BUS, NULL);
1048
1049
1050
1051

    for (c = soundhw; c->name; ++c) {
        if (c->enabled) {
            if (c->isa) {
1052
1053
1054
                if (!isa_bus) {
                    fprintf(stderr, "ISA bus not available for %s\n", c->name);
                    exit(1);
1055
                }
1056
                c->init.init_isa(isa_bus);
1057
            } else {
1058
1059
1060
                if (!pci_bus) {
                    fprintf(stderr, "PCI bus not available for %s\n", c->name);
                    exit(1);
1061
                }
1062
                c->init.init_pci(pci_bus);
1063
1064
1065
1066
            }
        }
    }
}
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084

int qemu_uuid_parse(const char *str, uint8_t *uuid)
{
    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;
    }
#ifdef TARGET_I386
1085
    smbios_add_field(1, offsetof(struct smbios_type_1, uuid), uuid, 16);
1086
1087
1088
1089
#endif
    return 0;
}

1090
void do_acpitable_option(const QemuOpts *opts)
1091
1092
{
#ifdef TARGET_I386
1093
1094
1095
1096
1097
1098
1099
    Error *err = NULL;

    acpi_table_add(opts, &err);
    if (err) {
        fprintf(stderr, "Wrong acpi table provided: %s\n",
                error_get_pretty(err));
        error_free(err);
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
        exit(1);
    }
#endif
}

void do_smbios_option(const char *optarg)
{
#ifdef TARGET_I386
    if (smbios_entry_add(optarg) < 0) {
        exit(1);
    }
#endif
}

void cpudef_init(void)
{
#if defined(cpudef_setup)
    cpudef_setup(); /* parse cpu definitions in target config file */
#endif
}

1121
1122
1123
1124
1125
int tcg_available(void)
{
    return 1;
}

Blue Swirl's avatar