dump.c 59.9 KB
Newer Older
1 2 3 4 5 6 7 8
/*
 * QEMU dump
 *
 * Copyright Fujitsu, Corp. 2011, 2012
 *
 * Authors:
 *     Wen Congyang <wency@cn.fujitsu.com>
 *
9 10
 * This work is licensed under the terms of the GNU GPL, version 2 or later.
 * See the COPYING file in the top-level directory.
11 12 13
 *
 */

Peter Maydell's avatar
Peter Maydell committed
14
#include "qemu/osdep.h"
15
#include "qemu/cutils.h"
16 17
#include "elf.h"
#include "cpu.h"
18
#include "exec/hwaddr.h"
19
#include "monitor/monitor.h"
20 21 22 23
#include "sysemu/kvm.h"
#include "sysemu/dump.h"
#include "sysemu/sysemu.h"
#include "sysemu/memory_mapping.h"
24
#include "sysemu/cpus.h"
25
#include "qapi/error.h"
26 27
#include "qapi/qapi-commands-misc.h"
#include "qapi/qapi-events-misc.h"
28
#include "qapi/qmp/qerror.h"
29 30
#include "qemu/error-report.h"
#include "hw/misc/vmcoreinfo.h"
31

32 33 34 35 36 37 38
#include <zlib.h>
#ifdef CONFIG_LZO
#include <lzo/lzo1x.h>
#endif
#ifdef CONFIG_SNAPPY
#include <snappy-c.h>
#endif
39 40 41
#ifndef ELF_MACHINE_UNAME
#define ELF_MACHINE_UNAME "Unknown"
#endif
42

43 44 45 46 47 48 49
#define MAX_GUEST_NOTE_SIZE (1 << 20) /* 1MB should be enough */

#define ELF_NOTE_SIZE(hdr_size, name_size, desc_size)   \
    ((DIV_ROUND_UP((hdr_size), 4) +                     \
      DIV_ROUND_UP((name_size), 4) +                    \
      DIV_ROUND_UP((desc_size), 4)) * 4)

50
uint16_t cpu_to_dump16(DumpState *s, uint16_t val)
51
{
52
    if (s->dump_info.d_endian == ELFDATA2LSB) {
53 54 55 56 57 58 59 60
        val = cpu_to_le16(val);
    } else {
        val = cpu_to_be16(val);
    }

    return val;
}

61
uint32_t cpu_to_dump32(DumpState *s, uint32_t val)
62
{
63
    if (s->dump_info.d_endian == ELFDATA2LSB) {
64 65 66 67 68 69 70 71
        val = cpu_to_le32(val);
    } else {
        val = cpu_to_be32(val);
    }

    return val;
}

72
uint64_t cpu_to_dump64(DumpState *s, uint64_t val)
73
{
74
    if (s->dump_info.d_endian == ELFDATA2LSB) {
75 76 77 78 79 80 81 82 83 84
        val = cpu_to_le64(val);
    } else {
        val = cpu_to_be64(val);
    }

    return val;
}

static int dump_cleanup(DumpState *s)
{
85
    guest_phys_blocks_free(&s->guest_phys_blocks);
86
    memory_mapping_list_free(&s->list);
87
    close(s->fd);
88 89
    g_free(s->guest_note);
    s->guest_note = NULL;
90
    if (s->resume) {
91 92 93
        if (s->detached) {
            qemu_mutex_lock_iothread();
        }
94
        vm_start();
95 96 97
        if (s->detached) {
            qemu_mutex_unlock_iothread();
        }
98 99
    }

100
    return 0;
101 102
}

103
static int fd_write_vmcore(const void *buf, size_t size, void *opaque)
104 105
{
    DumpState *s = opaque;
106 107 108 109
    size_t written_size;

    written_size = qemu_write_full(s->fd, buf, size);
    if (written_size != size) {
110
        return -errno;
111 112 113 114 115
    }

    return 0;
}

116
static void write_elf64_header(DumpState *s, Error **errp)
117 118 119 120 121 122 123 124 125
{
    Elf64_Ehdr elf_header;
    int ret;

    memset(&elf_header, 0, sizeof(Elf64_Ehdr));
    memcpy(&elf_header, ELFMAG, SELFMAG);
    elf_header.e_ident[EI_CLASS] = ELFCLASS64;
    elf_header.e_ident[EI_DATA] = s->dump_info.d_endian;
    elf_header.e_ident[EI_VERSION] = EV_CURRENT;
126 127 128 129 130 131 132
    elf_header.e_type = cpu_to_dump16(s, ET_CORE);
    elf_header.e_machine = cpu_to_dump16(s, s->dump_info.d_machine);
    elf_header.e_version = cpu_to_dump32(s, EV_CURRENT);
    elf_header.e_ehsize = cpu_to_dump16(s, sizeof(elf_header));
    elf_header.e_phoff = cpu_to_dump64(s, sizeof(Elf64_Ehdr));
    elf_header.e_phentsize = cpu_to_dump16(s, sizeof(Elf64_Phdr));
    elf_header.e_phnum = cpu_to_dump16(s, s->phdr_num);
133 134 135
    if (s->have_section) {
        uint64_t shoff = sizeof(Elf64_Ehdr) + sizeof(Elf64_Phdr) * s->sh_info;

136 137 138
        elf_header.e_shoff = cpu_to_dump64(s, shoff);
        elf_header.e_shentsize = cpu_to_dump16(s, sizeof(Elf64_Shdr));
        elf_header.e_shnum = cpu_to_dump16(s, 1);
139 140 141 142
    }

    ret = fd_write_vmcore(&elf_header, sizeof(elf_header), s);
    if (ret < 0) {
143
        error_setg_errno(errp, -ret, "dump: failed to write elf header");
144 145 146
    }
}

147
static void write_elf32_header(DumpState *s, Error **errp)
148 149 150 151 152 153 154
{
    Elf32_Ehdr elf_header;
    int ret;

    memset(&elf_header, 0, sizeof(Elf32_Ehdr));
    memcpy(&elf_header, ELFMAG, SELFMAG);
    elf_header.e_ident[EI_CLASS] = ELFCLASS32;
155
    elf_header.e_ident[EI_DATA] = s->dump_info.d_endian;
156
    elf_header.e_ident[EI_VERSION] = EV_CURRENT;
157 158 159 160 161 162 163
    elf_header.e_type = cpu_to_dump16(s, ET_CORE);
    elf_header.e_machine = cpu_to_dump16(s, s->dump_info.d_machine);
    elf_header.e_version = cpu_to_dump32(s, EV_CURRENT);
    elf_header.e_ehsize = cpu_to_dump16(s, sizeof(elf_header));
    elf_header.e_phoff = cpu_to_dump32(s, sizeof(Elf32_Ehdr));
    elf_header.e_phentsize = cpu_to_dump16(s, sizeof(Elf32_Phdr));
    elf_header.e_phnum = cpu_to_dump16(s, s->phdr_num);
164 165 166
    if (s->have_section) {
        uint32_t shoff = sizeof(Elf32_Ehdr) + sizeof(Elf32_Phdr) * s->sh_info;

167 168 169
        elf_header.e_shoff = cpu_to_dump32(s, shoff);
        elf_header.e_shentsize = cpu_to_dump16(s, sizeof(Elf32_Shdr));
        elf_header.e_shnum = cpu_to_dump16(s, 1);
170 171 172 173
    }

    ret = fd_write_vmcore(&elf_header, sizeof(elf_header), s);
    if (ret < 0) {
174
        error_setg_errno(errp, -ret, "dump: failed to write elf header");
175 176 177
    }
}

178 179 180
static void write_elf64_load(DumpState *s, MemoryMapping *memory_mapping,
                             int phdr_index, hwaddr offset,
                             hwaddr filesz, Error **errp)
181 182 183 184 185
{
    Elf64_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf64_Phdr));
186 187 188 189 190 191
    phdr.p_type = cpu_to_dump32(s, PT_LOAD);
    phdr.p_offset = cpu_to_dump64(s, offset);
    phdr.p_paddr = cpu_to_dump64(s, memory_mapping->phys_addr);
    phdr.p_filesz = cpu_to_dump64(s, filesz);
    phdr.p_memsz = cpu_to_dump64(s, memory_mapping->length);
    phdr.p_vaddr = cpu_to_dump64(s, memory_mapping->virt_addr);
192

193 194
    assert(memory_mapping->length >= filesz);

195 196
    ret = fd_write_vmcore(&phdr, sizeof(Elf64_Phdr), s);
    if (ret < 0) {
197 198
        error_setg_errno(errp, -ret,
                         "dump: failed to write program header table");
199 200 201
    }
}

202 203 204
static void write_elf32_load(DumpState *s, MemoryMapping *memory_mapping,
                             int phdr_index, hwaddr offset,
                             hwaddr filesz, Error **errp)
205 206 207 208 209
{
    Elf32_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf32_Phdr));
210 211 212 213 214 215
    phdr.p_type = cpu_to_dump32(s, PT_LOAD);
    phdr.p_offset = cpu_to_dump32(s, offset);
    phdr.p_paddr = cpu_to_dump32(s, memory_mapping->phys_addr);
    phdr.p_filesz = cpu_to_dump32(s, filesz);
    phdr.p_memsz = cpu_to_dump32(s, memory_mapping->length);
    phdr.p_vaddr = cpu_to_dump32(s, memory_mapping->virt_addr);
216

217 218
    assert(memory_mapping->length >= filesz);

219 220
    ret = fd_write_vmcore(&phdr, sizeof(Elf32_Phdr), s);
    if (ret < 0) {
221 222
        error_setg_errno(errp, -ret,
                         "dump: failed to write program header table");
223 224 225
    }
}

226
static void write_elf64_note(DumpState *s, Error **errp)
227 228
{
    Elf64_Phdr phdr;
229
    hwaddr begin = s->memory_offset - s->note_size;
230 231 232
    int ret;

    memset(&phdr, 0, sizeof(Elf64_Phdr));
233 234
    phdr.p_type = cpu_to_dump32(s, PT_NOTE);
    phdr.p_offset = cpu_to_dump64(s, begin);
235
    phdr.p_paddr = 0;
236 237
    phdr.p_filesz = cpu_to_dump64(s, s->note_size);
    phdr.p_memsz = cpu_to_dump64(s, s->note_size);
238 239 240 241
    phdr.p_vaddr = 0;

    ret = fd_write_vmcore(&phdr, sizeof(Elf64_Phdr), s);
    if (ret < 0) {
242 243
        error_setg_errno(errp, -ret,
                         "dump: failed to write program header table");
244 245 246
    }
}

247 248 249 250 251
static inline int cpu_index(CPUState *cpu)
{
    return cpu->cpu_index + 1;
}

252 253 254 255 256 257 258 259 260 261 262 263 264
static void write_guest_note(WriteCoreDumpFunction f, DumpState *s,
                             Error **errp)
{
    int ret;

    if (s->guest_note) {
        ret = f(s->guest_note, s->guest_note_size, s);
        if (ret < 0) {
            error_setg(errp, "dump: failed to write guest note");
        }
    }
}

265 266
static void write_elf64_notes(WriteCoreDumpFunction f, DumpState *s,
                              Error **errp)
267
{
268
    CPUState *cpu;
269 270 271
    int ret;
    int id;

272
    CPU_FOREACH(cpu) {
273
        id = cpu_index(cpu);
274
        ret = cpu_write_elf64_note(f, cpu, id, s);
275
        if (ret < 0) {
276
            error_setg(errp, "dump: failed to write elf notes");
277
            return;
278 279 280
        }
    }

281
    CPU_FOREACH(cpu) {
282
        ret = cpu_write_elf64_qemunote(f, cpu, s);
283
        if (ret < 0) {
284
            error_setg(errp, "dump: failed to write CPU status");
285
            return;
286 287
        }
    }
288 289

    write_guest_note(f, s, errp);
290 291
}

292
static void write_elf32_note(DumpState *s, Error **errp)
293
{
294
    hwaddr begin = s->memory_offset - s->note_size;
295 296 297 298
    Elf32_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf32_Phdr));
299 300
    phdr.p_type = cpu_to_dump32(s, PT_NOTE);
    phdr.p_offset = cpu_to_dump32(s, begin);
301
    phdr.p_paddr = 0;
302 303
    phdr.p_filesz = cpu_to_dump32(s, s->note_size);
    phdr.p_memsz = cpu_to_dump32(s, s->note_size);
304 305 306 307
    phdr.p_vaddr = 0;

    ret = fd_write_vmcore(&phdr, sizeof(Elf32_Phdr), s);
    if (ret < 0) {
308 309
        error_setg_errno(errp, -ret,
                         "dump: failed to write program header table");
310 311 312
    }
}

313 314
static void write_elf32_notes(WriteCoreDumpFunction f, DumpState *s,
                              Error **errp)
315
{
316
    CPUState *cpu;
317 318 319
    int ret;
    int id;

320
    CPU_FOREACH(cpu) {
321
        id = cpu_index(cpu);
322
        ret = cpu_write_elf32_note(f, cpu, id, s);
323
        if (ret < 0) {
324
            error_setg(errp, "dump: failed to write elf notes");
325
            return;
326 327 328
        }
    }

329
    CPU_FOREACH(cpu) {
330
        ret = cpu_write_elf32_qemunote(f, cpu, s);
331
        if (ret < 0) {
332
            error_setg(errp, "dump: failed to write CPU status");
333
            return;
334 335
        }
    }
336 337

    write_guest_note(f, s, errp);
338 339
}

340
static void write_elf_section(DumpState *s, int type, Error **errp)
341 342 343 344 345 346 347 348 349 350
{
    Elf32_Shdr shdr32;
    Elf64_Shdr shdr64;
    int shdr_size;
    void *shdr;
    int ret;

    if (type == 0) {
        shdr_size = sizeof(Elf32_Shdr);
        memset(&shdr32, 0, shdr_size);
351
        shdr32.sh_info = cpu_to_dump32(s, s->sh_info);
352 353 354 355
        shdr = &shdr32;
    } else {
        shdr_size = sizeof(Elf64_Shdr);
        memset(&shdr64, 0, shdr_size);
356
        shdr64.sh_info = cpu_to_dump32(s, s->sh_info);
357 358 359 360 361
        shdr = &shdr64;
    }

    ret = fd_write_vmcore(&shdr, shdr_size, s);
    if (ret < 0) {
362 363
        error_setg_errno(errp, -ret,
                         "dump: failed to write section header table");
364 365 366
    }
}

367
static void write_data(DumpState *s, void *buf, int length, Error **errp)
368 369 370 371 372
{
    int ret;

    ret = fd_write_vmcore(buf, length, s);
    if (ret < 0) {
373
        error_setg_errno(errp, -ret, "dump: failed to save memory");
374 375
    } else {
        s->written_size += length;
376 377 378
    }
}

379 380 381
/* write the memory to vmcore. 1 page per I/O. */
static void write_memory(DumpState *s, GuestPhysBlock *block, ram_addr_t start,
                         int64_t size, Error **errp)
382 383
{
    int64_t i;
384
    Error *local_err = NULL;
385

386 387 388
    for (i = 0; i < size / s->dump_info.page_size; i++) {
        write_data(s, block->host_addr + start + i * s->dump_info.page_size,
                   s->dump_info.page_size, &local_err);
389 390 391
        if (local_err) {
            error_propagate(errp, local_err);
            return;
392 393 394
        }
    }

395 396 397
    if ((size % s->dump_info.page_size) != 0) {
        write_data(s, block->host_addr + start + i * s->dump_info.page_size,
                   size % s->dump_info.page_size, &local_err);
398 399 400
        if (local_err) {
            error_propagate(errp, local_err);
            return;
401 402 403 404
        }
    }
}

405 406 407 408 409 410
/* get the memory's offset and size in the vmcore */
static void get_offset_range(hwaddr phys_addr,
                             ram_addr_t mapping_length,
                             DumpState *s,
                             hwaddr *p_offset,
                             hwaddr *p_filesz)
411
{
412
    GuestPhysBlock *block;
413
    hwaddr offset = s->memory_offset;
414 415
    int64_t size_in_block, start;

416 417 418 419
    /* When the memory is not stored into vmcore, offset will be -1 */
    *p_offset = -1;
    *p_filesz = 0;

420 421
    if (s->has_filter) {
        if (phys_addr < s->begin || phys_addr >= s->begin + s->length) {
422
            return;
423 424 425
        }
    }

426
    QTAILQ_FOREACH(block, &s->guest_phys_blocks.head, next) {
427
        if (s->has_filter) {
428 429
            if (block->target_start >= s->begin + s->length ||
                block->target_end <= s->begin) {
430 431 432 433
                /* This block is out of the range */
                continue;
            }

434 435
            if (s->begin <= block->target_start) {
                start = block->target_start;
436 437 438 439
            } else {
                start = s->begin;
            }

440 441 442
            size_in_block = block->target_end - start;
            if (s->begin + s->length < block->target_end) {
                size_in_block -= block->target_end - (s->begin + s->length);
443 444
            }
        } else {
445 446
            start = block->target_start;
            size_in_block = block->target_end - block->target_start;
447 448 449
        }

        if (phys_addr >= start && phys_addr < start + size_in_block) {
450 451 452
            *p_offset = phys_addr - start + offset;

            /* The offset range mapped from the vmcore file must not spill over
453
             * the GuestPhysBlock, clamp it. The rest of the mapping will be
454 455 456 457 458 459 460
             * zero-filled in memory at load time; see
             * <http://refspecs.linuxbase.org/elf/gabi4+/ch5.pheader.html>.
             */
            *p_filesz = phys_addr + mapping_length <= start + size_in_block ?
                        mapping_length :
                        size_in_block - (phys_addr - start);
            return;
461 462 463 464 465 466
        }

        offset += size_in_block;
    }
}

467
static void write_elf_loads(DumpState *s, Error **errp)
468
{
469
    hwaddr offset, filesz;
470 471 472
    MemoryMapping *memory_mapping;
    uint32_t phdr_index = 1;
    uint32_t max_index;
473
    Error *local_err = NULL;
474 475 476 477 478 479 480 481

    if (s->have_section) {
        max_index = s->sh_info;
    } else {
        max_index = s->phdr_num;
    }

    QTAILQ_FOREACH(memory_mapping, &s->list.head, next) {
482 483 484
        get_offset_range(memory_mapping->phys_addr,
                         memory_mapping->length,
                         s, &offset, &filesz);
485
        if (s->dump_info.d_class == ELFCLASS64) {
486 487
            write_elf64_load(s, memory_mapping, phdr_index++, offset,
                             filesz, &local_err);
488
        } else {
489 490
            write_elf32_load(s, memory_mapping, phdr_index++, offset,
                             filesz, &local_err);
491 492
        }

493 494 495
        if (local_err) {
            error_propagate(errp, local_err);
            return;
496 497 498 499 500 501 502 503 504
        }

        if (phdr_index >= max_index) {
            break;
        }
    }
}

/* write elf header, PT_NOTE and elf note to vmcore. */
505
static void dump_begin(DumpState *s, Error **errp)
506
{
507
    Error *local_err = NULL;
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534

    /*
     * the vmcore's format is:
     *   --------------
     *   |  elf header |
     *   --------------
     *   |  PT_NOTE    |
     *   --------------
     *   |  PT_LOAD    |
     *   --------------
     *   |  ......     |
     *   --------------
     *   |  PT_LOAD    |
     *   --------------
     *   |  sec_hdr    |
     *   --------------
     *   |  elf note   |
     *   --------------
     *   |  memory     |
     *   --------------
     *
     * we only know where the memory is saved after we write elf note into
     * vmcore.
     */

    /* write elf header to vmcore */
    if (s->dump_info.d_class == ELFCLASS64) {
535
        write_elf64_header(s, &local_err);
536
    } else {
537
        write_elf32_header(s, &local_err);
538
    }
539 540 541
    if (local_err) {
        error_propagate(errp, local_err);
        return;
542 543 544 545
    }

    if (s->dump_info.d_class == ELFCLASS64) {
        /* write PT_NOTE to vmcore */
546 547 548 549
        write_elf64_note(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
550 551 552
        }

        /* write all PT_LOAD to vmcore */
553 554 555 556
        write_elf_loads(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
557 558 559 560
        }

        /* write section to vmcore */
        if (s->have_section) {
561 562 563 564
            write_elf_section(s, 1, &local_err);
            if (local_err) {
                error_propagate(errp, local_err);
                return;
565 566 567 568
            }
        }

        /* write notes to vmcore */
569 570 571 572
        write_elf64_notes(fd_write_vmcore, s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
573 574 575
        }
    } else {
        /* write PT_NOTE to vmcore */
576 577 578 579
        write_elf32_note(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
580 581 582
        }

        /* write all PT_LOAD to vmcore */
583 584 585 586
        write_elf_loads(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
587 588 589 590
        }

        /* write section to vmcore */
        if (s->have_section) {
591 592 593 594
            write_elf_section(s, 0, &local_err);
            if (local_err) {
                error_propagate(errp, local_err);
                return;
595 596 597 598
            }
        }

        /* write notes to vmcore */
599 600 601 602
        write_elf32_notes(fd_write_vmcore, s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
603 604 605 606
        }
    }
}

607
static int get_next_block(DumpState *s, GuestPhysBlock *block)
608 609
{
    while (1) {
610
        block = QTAILQ_NEXT(block, next);
611 612 613 614 615 616
        if (!block) {
            /* no more block */
            return 1;
        }

        s->start = 0;
617
        s->next_block = block;
618
        if (s->has_filter) {
619 620
            if (block->target_start >= s->begin + s->length ||
                block->target_end <= s->begin) {
621 622 623 624
                /* This block is out of the range */
                continue;
            }

625 626
            if (s->begin > block->target_start) {
                s->start = s->begin - block->target_start;
627 628 629 630 631 632 633 634
            }
        }

        return 0;
    }
}

/* write all memory to vmcore */
635
static void dump_iterate(DumpState *s, Error **errp)
636
{
637
    GuestPhysBlock *block;
638
    int64_t size;
639
    Error *local_err = NULL;
640

641
    do {
642
        block = s->next_block;
643

644
        size = block->target_end - block->target_start;
645 646
        if (s->has_filter) {
            size -= s->start;
647 648
            if (s->begin + s->length < block->target_end) {
                size -= block->target_end - (s->begin + s->length);
649 650
            }
        }
651 652 653 654
        write_memory(s, block, s->start, size, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
655 656
        }

657
    } while (!get_next_block(s, block));
658 659
}

660
static void create_vmcore(DumpState *s, Error **errp)
661
{
662
    Error *local_err = NULL;
663

664 665 666 667
    dump_begin(s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
668 669
    }

670
    dump_iterate(s, errp);
671 672
}

673 674
static int write_start_flat_header(int fd)
{
675
    MakedumpfileHeader *mh;
676 677
    int ret = 0;

678 679
    QEMU_BUILD_BUG_ON(sizeof *mh > MAX_SIZE_MDF_HEADER);
    mh = g_malloc0(MAX_SIZE_MDF_HEADER);
680

681 682
    memcpy(mh->signature, MAKEDUMPFILE_SIGNATURE,
           MIN(sizeof mh->signature, sizeof MAKEDUMPFILE_SIGNATURE));
683

684 685
    mh->type = cpu_to_be64(TYPE_FLAT_HEADER);
    mh->version = cpu_to_be64(VERSION_FLAT_HEADER);
686 687

    size_t written_size;
688
    written_size = qemu_write_full(fd, mh, MAX_SIZE_MDF_HEADER);
689 690 691 692
    if (written_size != MAX_SIZE_MDF_HEADER) {
        ret = -1;
    }

693
    g_free(mh);
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
    return ret;
}

static int write_end_flat_header(int fd)
{
    MakedumpfileDataHeader mdh;

    mdh.offset = END_FLAG_FLAT_HEADER;
    mdh.buf_size = END_FLAG_FLAT_HEADER;

    size_t written_size;
    written_size = qemu_write_full(fd, &mdh, sizeof(mdh));
    if (written_size != sizeof(mdh)) {
        return -1;
    }

    return 0;
}

713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
static int write_buffer(int fd, off_t offset, const void *buf, size_t size)
{
    size_t written_size;
    MakedumpfileDataHeader mdh;

    mdh.offset = cpu_to_be64(offset);
    mdh.buf_size = cpu_to_be64(size);

    written_size = qemu_write_full(fd, &mdh, sizeof(mdh));
    if (written_size != sizeof(mdh)) {
        return -1;
    }

    written_size = qemu_write_full(fd, buf, size);
    if (written_size != size) {
        return -1;
    }

    return 0;
}

734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
static int buf_write_note(const void *buf, size_t size, void *opaque)
{
    DumpState *s = opaque;

    /* note_buf is not enough */
    if (s->note_buf_offset + size > s->note_size) {
        return -1;
    }

    memcpy(s->note_buf + s->note_buf_offset, buf, size);

    s->note_buf_offset += size;

    return 0;
}

750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787
/*
 * This function retrieves various sizes from an elf header.
 *
 * @note has to be a valid ELF note. The return sizes are unmodified
 * (not padded or rounded up to be multiple of 4).
 */
static void get_note_sizes(DumpState *s, const void *note,
                           uint64_t *note_head_size,
                           uint64_t *name_size,
                           uint64_t *desc_size)
{
    uint64_t note_head_sz;
    uint64_t name_sz;
    uint64_t desc_sz;

    if (s->dump_info.d_class == ELFCLASS64) {
        const Elf64_Nhdr *hdr = note;
        note_head_sz = sizeof(Elf64_Nhdr);
        name_sz = tswap64(hdr->n_namesz);
        desc_sz = tswap64(hdr->n_descsz);
    } else {
        const Elf32_Nhdr *hdr = note;
        note_head_sz = sizeof(Elf32_Nhdr);
        name_sz = tswap32(hdr->n_namesz);
        desc_sz = tswap32(hdr->n_descsz);
    }

    if (note_head_size) {
        *note_head_size = note_head_sz;
    }
    if (name_size) {
        *name_size = name_sz;
    }
    if (desc_size) {
        *desc_size = desc_sz;
    }
}

788 789 790 791 792 793 794 795 796
static bool note_name_equal(DumpState *s,
                            const uint8_t *note, const char *name)
{
    int len = strlen(name) + 1;
    uint64_t head_size, name_size;

    get_note_sizes(s, note, &head_size, &name_size, NULL);
    head_size = ROUND_UP(head_size, 4);

797
    return name_size == len && memcmp(note + head_size, name, len) == 0;
798 799
}

800
/* write common header, sub header and elf note to vmcore */
801
static void create_header32(DumpState *s, Error **errp)
802 803 804 805 806 807 808 809 810
{
    DiskDumpHeader32 *dh = NULL;
    KdumpSubHeader32 *kh = NULL;
    size_t size;
    uint32_t block_size;
    uint32_t sub_hdr_size;
    uint32_t bitmap_blocks;
    uint32_t status = 0;
    uint64_t offset_note;
811
    Error *local_err = NULL;
812 813 814 815 816

    /* write common header, the version of kdump-compressed format is 6th */
    size = sizeof(DiskDumpHeader32);
    dh = g_malloc0(size);

817
    memcpy(dh->signature, KDUMP_SIGNATURE, SIG_LEN);
818
    dh->header_version = cpu_to_dump32(s, 6);
819
    block_size = s->dump_info.page_size;
820
    dh->block_size = cpu_to_dump32(s, block_size);
821 822
    sub_hdr_size = sizeof(struct KdumpSubHeader32) + s->note_size;
    sub_hdr_size = DIV_ROUND_UP(sub_hdr_size, block_size);
823
    dh->sub_hdr_size = cpu_to_dump32(s, sub_hdr_size);
824
    /* dh->max_mapnr may be truncated, full 64bit is in kh.max_mapnr_64 */
825 826
    dh->max_mapnr = cpu_to_dump32(s, MIN(s->max_mapnr, UINT_MAX));
    dh->nr_cpus = cpu_to_dump32(s, s->nr_cpus);
827
    bitmap_blocks = DIV_ROUND_UP(s->len_dump_bitmap, block_size) * 2;
828
    dh->bitmap_blocks = cpu_to_dump32(s, bitmap_blocks);
829
    strncpy(dh->utsname.machine, ELF_MACHINE_UNAME, sizeof(dh->utsname.machine));
830 831 832 833 834 835 836 837 838 839 840 841 842 843

    if (s->flag_compress & DUMP_DH_COMPRESSED_ZLIB) {
        status |= DUMP_DH_COMPRESSED_ZLIB;
    }
#ifdef CONFIG_LZO
    if (s->flag_compress & DUMP_DH_COMPRESSED_LZO) {
        status |= DUMP_DH_COMPRESSED_LZO;
    }
#endif
#ifdef CONFIG_SNAPPY
    if (s->flag_compress & DUMP_DH_COMPRESSED_SNAPPY) {
        status |= DUMP_DH_COMPRESSED_SNAPPY;
    }
#endif
844
    dh->status = cpu_to_dump32(s, status);
845 846

    if (write_buffer(s->fd, 0, dh, size) < 0) {
847
        error_setg(errp, "dump: failed to write disk dump header");
848 849 850 851 852 853 854 855
        goto out;
    }

    /* write sub header */
    size = sizeof(KdumpSubHeader32);
    kh = g_malloc0(size);

    /* 64bit max_mapnr_64 */
856
    kh->max_mapnr_64 = cpu_to_dump64(s, s->max_mapnr);
857
    kh->phys_base = cpu_to_dump32(s, s->dump_info.phys_base);
858
    kh->dump_level = cpu_to_dump32(s, DUMP_LEVEL);
859 860

    offset_note = DISKDUMP_HEADER_BLOCKS * block_size + size;
861 862 863 864 865 866 867 868 869 870 871 872
    if (s->guest_note &&
        note_name_equal(s, s->guest_note, "VMCOREINFO")) {
        uint64_t hsize, name_size, size_vmcoreinfo_desc, offset_vmcoreinfo;

        get_note_sizes(s, s->guest_note,
                       &hsize, &name_size, &size_vmcoreinfo_desc);
        offset_vmcoreinfo = offset_note + s->note_size - s->guest_note_size +
            (DIV_ROUND_UP(hsize, 4) + DIV_ROUND_UP(name_size, 4)) * 4;
        kh->offset_vmcoreinfo = cpu_to_dump64(s, offset_vmcoreinfo);
        kh->size_vmcoreinfo = cpu_to_dump32(s, size_vmcoreinfo_desc);
    }

873 874
    kh->offset_note = cpu_to_dump64(s, offset_note);
    kh->note_size = cpu_to_dump32(s, s->note_size);
875 876 877

    if (write_buffer(s->fd, DISKDUMP_HEADER_BLOCKS *
                     block_size, kh, size) < 0) {
878
        error_setg(errp, "dump: failed to write kdump sub header");
879 880 881 882 883 884 885 886
        goto out;
    }

    /* write note */
    s->note_buf = g_malloc0(s->note_size);
    s->note_buf_offset = 0;

    /* use s->note_buf to store notes temporarily */
887 888 889
    write_elf32_notes(buf_write_note, s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
890 891 892 893
        goto out;
    }
    if (write_buffer(s->fd, offset_note, s->note_buf,
                     s->note_size) < 0) {
894
        error_setg(errp, "dump: failed to write notes");
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
        goto out;
    }

    /* get offset of dump_bitmap */
    s->offset_dump_bitmap = (DISKDUMP_HEADER_BLOCKS + sub_hdr_size) *
                             block_size;

    /* get offset of page */
    s->offset_page = (DISKDUMP_HEADER_BLOCKS + sub_hdr_size + bitmap_blocks) *
                     block_size;

out:
    g_free(dh);
    g_free(kh);
    g_free(s->note_buf);
}

/* write common header, sub header and elf note to vmcore */
913
static void create_header64(DumpState *s, Error **errp)
914 915 916 917 918 919 920 921 922
{
    DiskDumpHeader64 *dh = NULL;
    KdumpSubHeader64 *kh = NULL;
    size_t size;
    uint32_t block_size;
    uint32_t sub_hdr_size;
    uint32_t bitmap_blocks;
    uint32_t status = 0;
    uint64_t offset_note;
923
    Error *local_err = NULL;
924 925 926 927 928

    /* write common header, the version of kdump-compressed format is 6th */
    size = sizeof(DiskDumpHeader64);
    dh = g_malloc0(size);

929
    memcpy(dh->signature, KDUMP_SIGNATURE, SIG_LEN);
930
    dh->header_version = cpu_to_dump32(s, 6);
931
    block_size = s->dump_info.page_size;
932
    dh->block_size = cpu_to_dump32(s, block_size);
933 934
    sub_hdr_size = sizeof(struct KdumpSubHeader64) + s->note_size;
    sub_hdr_size = DIV_ROUND_UP(sub_hdr_size, block_size);
935
    dh->sub_hdr_size = cpu_to_dump32(s, sub_hdr_size);
936
    /* dh->max_mapnr may be truncated, full 64bit is in kh.max_mapnr_64 */
937 938
    dh->max_mapnr = cpu_to_dump32(s, MIN(s->max_mapnr, UINT_MAX));
    dh->nr_cpus = cpu_to_dump32(s, s->nr_cpus);
939
    bitmap_blocks = DIV_ROUND_UP(s->len_dump_bitmap, block_size) * 2;
940
    dh->bitmap_blocks = cpu_to_dump32(s, bitmap_blocks);
941
    strncpy(dh->utsname.machine, ELF_MACHINE_UNAME, sizeof(dh->utsname.machine));
942 943 944 945 946 947 948 949 950 951 952 953 954 955

    if (s->flag_compress & DUMP_DH_COMPRESSED_ZLIB) {
        status |= DUMP_DH_COMPRESSED_ZLIB;
    }
#ifdef CONFIG_LZO
    if (s->flag_compress & DUMP_DH_COMPRESSED_LZO) {
        status |= DUMP_DH_COMPRESSED_LZO;
    }
#endif
#ifdef CONFIG_SNAPPY
    if (s->flag_compress & DUMP_DH_COMPRESSED_SNAPPY) {
        status |= DUMP_DH_COMPRESSED_SNAPPY;
    }
#endif
956
    dh->status = cpu_to_dump32(s, status);
957 958

    if (write_buffer(s->fd, 0, dh, size) < 0) {
959
        error_setg(errp, "dump: failed to write disk dump header");
960 961 962 963 964 965 966 967
        goto out;
    }

    /* write sub header */
    size = sizeof(KdumpSubHeader64);
    kh = g_malloc0(size);

    /* 64bit max_mapnr_64 */
968
    kh->max_mapnr_64 = cpu_to_dump64(s, s->max_mapnr);
969
    kh->phys_base = cpu_to_dump64(s, s->dump_info.phys_base);
970
    kh->dump_level = cpu_to_dump32(s, DUMP_LEVEL);
971 972

    offset_note = DISKDUMP_HEADER_BLOCKS * block_size + size;
973 974 975 976 977 978 979 980 981 982 983 984
    if (s->guest_note &&
        note_name_equal(s, s->guest_note, "VMCOREINFO")) {
        uint64_t hsize, name_size, size_vmcoreinfo_desc, offset_vmcoreinfo;

        get_note_sizes(s, s->guest_note,
                       &hsize, &name_size, &size_vmcoreinfo_desc);
        offset_vmcoreinfo = offset_note + s->note_size - s->guest_note_size +
            (DIV_ROUND_UP(hsize, 4) + DIV_ROUND_UP(name_size, 4)) * 4;
        kh->offset_vmcoreinfo = cpu_to_dump64(s, offset_vmcoreinfo);
        kh->size_vmcoreinfo = cpu_to_dump64(s, size_vmcoreinfo_desc);
    }

985 986
    kh->offset_note = cpu_to_dump64(s, offset_note);
    kh->note_size = cpu_to_dump64(s, s->note_size);
987 988 989

    if (write_buffer(s->fd, DISKDUMP_HEADER_BLOCKS *
                     block_size, kh, size) < 0) {
990
        error_setg(errp, "dump: failed to write kdump sub header");
991 992 993 994 995 996 997 998
        goto out;
    }

    /* write note */
    s->note_buf = g_malloc0(s->note_size);
    s->note_buf_offset = 0;

    /* use s->note_buf to store notes temporarily */
999 1000 1001
    write_elf64_notes(buf_write_note, s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1002 1003 1004 1005 1006
        goto out;
    }

    if (write_buffer(s->fd, offset_note, s->note_buf,
                     s->note_size) < 0) {
1007
        error_setg(errp, "dump: failed to write notes");
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
        goto out;
    }

    /* get offset of dump_bitmap */
    s->offset_dump_bitmap = (DISKDUMP_HEADER_BLOCKS + sub_hdr_size) *
                             block_size;

    /* get offset of page */
    s->offset_page = (DISKDUMP_HEADER_BLOCKS + sub_hdr_size + bitmap_blocks) *
                     block_size;

out:
    g_free(dh);
    g_free(kh);
    g_free(s->note_buf);
}

1025
static void write_dump_header(DumpState *s, Error **errp)
1026
{
1027 1028
     Error *local_err = NULL;

1029
    if (s->dump_info.d_class == ELFCLASS32) {
1030
        create_header32(s, &local_err);
1031
    } else {
1032 1033
        create_header64(s, &local_err);
    }
1034
    error_propagate(errp, local_err);
1035 1036
}

1037 1038 1039 1040 1041
static size_t dump_bitmap_get_bufsize(DumpState *s)
{
    return s->dump_info.page_size;
}

1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
/*
 * set dump_bitmap sequencely. the bit before last_pfn is not allowed to be
 * rewritten, so if need to set the first bit, set last_pfn and pfn to 0.
 * set_dump_bitmap will always leave the recently set bit un-sync. And setting
 * (last bit + sizeof(buf) * 8) to 0 will do flushing the content in buf into
 * vmcore, ie. synchronizing un-sync bit into vmcore.
 */
static int set_dump_bitmap(uint64_t last_pfn, uint64_t pfn, bool value,
                           uint8_t *buf, DumpState *s)
{
    off_t old_offset, new_offset;
    off_t offset_bitmap1, offset_bitmap2;
    uint32_t byte, bit;
1055 1056
    size_t bitmap_bufsize = dump_bitmap_get_bufsize(s);
    size_t bits_per_buf = bitmap_bufsize * CHAR_BIT;
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066

    /* should not set the previous place */
    assert(last_pfn <= pfn);

    /*
     * if the bit needed to be set is not cached in buf, flush the data in buf
     * to vmcore firstly.
     * making new_offset be bigger than old_offset can also sync remained data
     * into vmcore.
     */
1067 1068
    old_offset = bitmap_bufsize * (last_pfn / bits_per_buf);
    new_offset = bitmap_bufsize * (pfn / bits_per_buf);
1069 1070 1071 1072 1073

    while (old_offset < new_offset) {
        /* calculate the offset and write dump_bitmap */
        offset_bitmap1 = s->offset_dump_bitmap + old_offset;
        if (write_buffer(s->fd, offset_bitmap1, buf,
1074
                         bitmap_bufsize) < 0) {
1075 1076 1077 1078 1079 1080 1081
            return -1;
        }

        /* dump level 1 is chosen, so 1st and 2nd bitmap are same */
        offset_bitmap2 = s->offset_dump_bitmap + s->len_dump_bitmap +
                         old_offset;
        if (write_buffer(s->fd, offset_bitmap2, buf,
1082
                         bitmap_bufsize) < 0) {
1083 1084 1085
            return -1;
        }

1086 1087
        memset(buf, 0, bitmap_bufsize);
        old_offset += bitmap_bufsize;
1088 1089 1090
    }

    /* get the exact place of the bit in the buf, and set it */
1091 1092
    byte = (pfn % bits_per_buf) / CHAR_BIT;
    bit = (pfn % bits_per_buf) % CHAR_BIT;
1093 1094 1095 1096 1097 1098 1099 1100 1101
    if (value) {
        buf[byte] |= 1u << bit;
    } else {
        buf[byte] &= ~(1u << bit);
    }

    return 0;
}

1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
static uint64_t dump_paddr_to_pfn(DumpState *s, uint64_t addr)
{
    int target_page_shift = ctz32(s->dump_info.page_size);

    return (addr >> target_page_shift) - ARCH_PFN_OFFSET;
}

static uint64_t dump_pfn_to_paddr(DumpState *s, uint64_t pfn)
{
    int target_page_shift = ctz32(s->dump_info.page_size);

    return (pfn + ARCH_PFN_OFFSET) << target_page_shift;
}

1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
/*
 * exam every page and return the page frame number and the address of the page.
 * bufptr can be NULL. note: the blocks here is supposed to reflect guest-phys
 * blocks, so block->target_start and block->target_end should be interal
 * multiples of the target page size.
 */
static bool get_next_page(GuestPhysBlock **blockptr, uint64_t *pfnptr,
                          uint8_t **bufptr, DumpState *s)
{
    GuestPhysBlock *block = *blockptr;
1126
    hwaddr addr, target_page_mask = ~((hwaddr)s->dump_info.page_size - 1);
1127 1128 1129 1130 1131 1132
    uint8_t *buf;

    /* block == NULL means the start of the iteration */
    if (!block) {
        block = QTAILQ_FIRST(&s->guest_phys_blocks.head);
        *blockptr = block;
1133 1134 1135
        assert((block->target_start & ~target_page_mask) == 0);
        assert((block->target_end & ~target_page_mask) == 0);
        *pfnptr = dump_paddr_to_pfn(s, block->target_start);
1136 1137 1138 1139 1140 1141 1142
        if (bufptr) {
            *bufptr = block->host_addr;
        }
        return true;
    }

    *pfnptr = *pfnptr + 1;
1143
    addr = dump_pfn_to_paddr(s, *pfnptr);
1144 1145

    if ((addr >= block->target_start) &&
1146
        (addr + s->dump_info.page_size <= block->target_end)) {
1147 1148 1149 1150 1151 1152 1153 1154
        buf = block->host_addr + (addr - block->target_start);
    } else {
        /* the next page is in the next block */
        block = QTAILQ_NEXT(block, next);
        *blockptr = block;
        if (!block) {
            return false;
        }
1155 1156 1157
        assert((block->target_start & ~target_page_mask) == 0);
        assert((block->target_end & ~target_page_mask) == 0);
        *pfnptr = dump_paddr_to_pfn(s, block->target_start);
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
        buf = block->host_addr;
    }

    if (bufptr) {
        *bufptr = buf;
    }

    return true;
}

1168
static void write_dump_bitmap(DumpState *s, Error **errp)
1169 1170 1171 1172 1173 1174
{
    int ret = 0;
    uint64_t last_pfn, pfn;
    void *dump_bitmap_buf;
    size_t num_dumpable;
    GuestPhysBlock *block_iter = NULL;
1175 1176
    size_t bitmap_bufsize = dump_bitmap_get_bufsize(s);
    size_t bits_per_buf = bitmap_bufsize * CHAR_BIT;
1177 1178

    /* dump_bitmap_buf is used to store dump_bitmap temporarily */
1179
    dump_bitmap_buf = g_malloc0(bitmap_bufsize);
1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190

    num_dumpable = 0;
    last_pfn = 0;

    /*
     * exam memory page by page, and set the bit in dump_bitmap corresponded
     * to the existing page.
     */
    while (get_next_page(&block_iter, &pfn, NULL, s)) {
        ret = set_dump_bitmap(last_pfn, pfn, true, dump_bitmap_buf, s);
        if (ret < 0) {
1191
            error_setg(errp, "dump: failed to set dump_bitmap");
1192 1193 1194 1195 1196 1197 1198 1199 1200
            goto out;
        }

        last_pfn = pfn;
        num_dumpable++;
    }

    /*
     * set_dump_bitmap will always leave the recently set bit un-sync. Here we
1201 1202
     * set the remaining bits from last_pfn to the end of the bitmap buffer to
     * 0. With those set, the un-sync bit will be synchronized into the vmcore.
1203 1204
     */
    if (num_dumpable > 0) {
1205
        ret = set_dump_bitmap(last_pfn, last_pfn + bits_per_buf, false,
1206 1207
                              dump_bitmap_buf, s);
        if (ret < 0) {
1208
            error_setg(errp, "dump: failed to sync dump_bitmap");
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
            goto out;
        }
    }

    /* number of dumpable pages that will be dumped later */
    s->num_dumpable = num_dumpable;

out:
    g_free(dump_bitmap_buf);
}

1220 1221 1222 1223 1224
static void prepare_data_cache(DataCache *data_cache, DumpState *s,
                               off_t offset)
{
    data_cache->fd = s->fd;
    data_cache->data_size = 0;
1225 1226
    data_cache->buf_size = 4 * dump_bitmap_get_bufsize(s);
    data_cache->buf = g_malloc0(data_cache->buf_size);
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
    data_cache->offset = offset;
}

static int write_cache(DataCache *dc, const void *buf, size_t size,
                       bool flag_sync)
{
    /*
     * dc->buf_size should not be less than size, otherwise dc will never be
     * enough
     */
    assert(size <= dc->buf_size);

    /*
     * if flag_sync is set, synchronize data in dc->buf into vmcore.
     * otherwise check if the space is enough for caching data in buf, if not,
     * write the data in dc->buf to dc->fd and reset dc->buf
     */
    if ((!flag_sync && dc->data_size + size > dc->buf_size) ||
        (flag_sync && dc->data_size > 0)) {
        if (write_buffer(dc->fd, dc->offset, dc->buf, dc->data_size) < 0) {
            return -1;
        }

        dc->offset += dc->data_size;
        dc->data_size = 0;
    }

    if (!flag_sync) {
        memcpy(dc->buf + dc->data_size, buf, size);
        dc->data_size += size;
    }

    return 0;
}

static void free_data_cache(DataCache *data_cache)
{
    g_free(data_cache->buf);
}

1267 1268
static size_t get_len_buf_out(size_t page_size, uint32_t flag_compress)
{
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279
    switch (flag_compress) {
    case DUMP_DH_COMPRESSED_ZLIB:
        return compressBound(page_size);

    case DUMP_DH_COMPRESSED_LZO:
        /*
         * LZO will expand incompressible data by a little amount. Please check
         * the following URL to see the expansion calculation:
         * http://www.oberhumer.com/opensource/lzo/lzofaq.php
         */
        return page_size + page_size / 16 + 64 + 3;
1280 1281

#ifdef CONFIG_SNAPPY
1282 1283
    case DUMP_DH_COMPRESSED_SNAPPY:
        return snappy_max_compressed_length(page_size);
1284
#endif
1285 1286
    }
    return 0;
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
}

/*
 * check if the page is all 0
 */
static inline bool is_zero_page(const uint8_t *buf, size_t page_size)
{
    return buffer_is_zero(buf, page_size);
}

1297
static void write_dump_pages(DumpState *s, Error **errp)
1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319
{
    int ret = 0;
    DataCache page_desc, page_data;
    size_t len_buf_out, size_out;
#ifdef CONFIG_LZO
    lzo_bytep wrkmem = NULL;
#endif
    uint8_t *buf_out = NULL;
    off_t offset_desc, offset_data;
    PageDescriptor pd, pd_zero;
    uint8_t *buf;
    GuestPhysBlock *block_iter = NULL;
    uint64_t pfn_iter;

    /* get offset of page_desc and page_data in dump file */
    offset_desc = s->offset_page;
    offset_data = offset_desc + sizeof(PageDescriptor) * s->num_dumpable;

    prepare_data_cache(&page_desc, s, offset_desc);
    prepare_data_cache(&page_data, s, offset_data);

    /* prepare buffer to store compressed data */
1320
    len_buf_out = get_len_buf_out(s->dump_info.page_size, s->flag_compress);
1321
    assert(len_buf_out != 0);
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332

#ifdef CONFIG_LZO
    wrkmem = g_malloc(LZO1X_1_MEM_COMPRESS);
#endif

    buf_out = g_malloc(len_buf_out);

    /*
     * init zero page's page_desc and page_data, because every zero page
     * uses the same page_data
     */
1333
    pd_zero.size = cpu_to_dump32(s, s->dump_info.page_size);
1334 1335 1336
    pd_zero.flags = cpu_to_dump32(s, 0);
    pd_zero.offset = cpu_to_dump64(s, offset_data);
    pd_zero.page_flags = cpu_to_dump64(s, 0);
1337 1338
    buf = g_malloc0(s->dump_info.page_size);
    ret = write_cache(&page_data, buf, s->dump_info.page_size, false);
1339 1340
    g_free(buf);
    if (ret < 0) {
1341
        error_setg(errp, "dump: failed to write page data (zero page)");
1342 1343 1344
        goto out;
    }

1345
    offset_data += s->dump_info.page_size;
1346 1347 1348 1349 1350 1351 1352

    /*
     * dump memory to vmcore page by page. zero page will all be resided in the
     * first page of page section
     */
    while (get_next_page(&block_iter, &pfn_iter, &buf, s)) {
        /* check zero page */
1353
        if (is_zero_page(buf, s->dump_info.page_size)) {
1354 1355 1356
            ret = write_cache(&page_desc, &pd_zero, sizeof(PageDescriptor),
                              false);
            if (ret < 0) {
1357
                error_setg(errp, "dump: failed to write page desc");
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
                goto out;
            }
        } else {
            /*
             * not zero page, then:
             * 1. compress the page
             * 2. write the compressed page into the cache of page_data
             * 3. get page desc of the compressed page and write it into the
             *    cache of page_desc
             *
             * only one compression format will be used here, for
             * s->flag_compress is set. But when compression fails to work,
             * we fall back to save in plaintext.
             */
             size_out = len_buf_out;
             if ((s->flag_compress & DUMP_DH_COMPRESSED_ZLIB) &&
1374
                    (compress2(buf_out, (uLongf *)&size_out, buf,
1375 1376
                               s->dump_info.page_size, Z_BEST_SPEED) == Z_OK) &&
                    (size_out < s->dump_info.page_size)) {
1377 1378
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_ZLIB);
                pd.size  = cpu_to_dump32(s, size_out);
1379 1380 1381

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1382
                    error_setg(errp, "dump: failed to write page data");
1383 1384 1385 1386
                    goto out;
                }
#ifdef CONFIG_LZO
            } else if ((s->flag_compress & DUMP_DH_COMPRESSED_LZO) &&
1387
                    (lzo1x_1_compress(buf, s->dump_info.page_size, buf_out,
1388
                    (lzo_uint *)&size_out, wrkmem) == LZO_E_OK) &&
1389
                    (size_out < s->dump_info.page_size)) {
1390 1391
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_LZO);
                pd.size  = cpu_to_dump32(s, size_out);
1392 1393 1394

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1395
                    error_setg(errp, "dump: failed to write page data");
1396 1397 1398 1399 1400
                    goto out;
                }
#endif
#ifdef CONFIG_SNAPPY
            } else if ((s->flag_compress & DUMP_DH_COMPRESSED_SNAPPY) &&
1401
                    (snappy_compress((char *)buf, s->dump_info.page_size,
1402
                    (char *)buf_out, &size_out) == SNAPPY_OK) &&
1403
                    (size_out < s->dump_info.page_size)) {
1404 1405
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_SNAPPY);
                pd.size  = cpu_to_dump32(s, size_out);
1406 1407 1408

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1409
                    error_setg(errp, "dump: failed to write page data");
1410 1411 1412 1413 1414 1415
                    goto out;
                }
#endif
            } else {
                /*
                 * fall back to save in plaintext, size_out should be
1416
                 * assigned the target's page size
1417
                 */
1418
                pd.flags = cpu_to_dump32(s, 0);
1419
                size_out = s->dump_info.page_size;
1420
                pd.size = cpu_to_dump32(s, size_out);
1421

1422 1423
                ret = write_cache(&page_data, buf,
                                  s->dump_info.page_size, false);
1424
                if (ret < 0) {
1425
                    error_setg(errp, "dump: failed to write page data");
1426 1427 1428 1429 1430
                    goto out;
                }
            }

            /* get and write page desc here */
1431 1432
            pd.page_flags = cpu_to_dump64(s, 0);
            pd.offset = cpu_to_dump64(s, offset_data);
1433 1434 1435 1436
            offset_data += size_out;

            ret = write_cache(&page_desc, &pd, sizeof