dump.c 59.4 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/qmp/qerror.h"
26
#include "qmp-commands.h"
27
#include "qapi-event.h"
28 29
#include "qemu/error-report.h"
#include "hw/misc/vmcoreinfo.h"
30

31 32 33 34 35 36 37
#include <zlib.h>
#ifdef CONFIG_LZO
#include <lzo/lzo1x.h>
#endif
#ifdef CONFIG_SNAPPY
#include <snappy-c.h>
#endif
38 39 40
#ifndef ELF_MACHINE_UNAME
#define ELF_MACHINE_UNAME "Unknown"
#endif
41

42 43 44 45 46 47 48
#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)

49
uint16_t cpu_to_dump16(DumpState *s, uint16_t val)
50
{
51
    if (s->dump_info.d_endian == ELFDATA2LSB) {
52 53 54 55 56 57 58 59
        val = cpu_to_le16(val);
    } else {
        val = cpu_to_be16(val);
    }

    return val;
}

60
uint32_t cpu_to_dump32(DumpState *s, uint32_t val)
61
{
62
    if (s->dump_info.d_endian == ELFDATA2LSB) {
63 64 65 66 67 68 69 70
        val = cpu_to_le32(val);
    } else {
        val = cpu_to_be32(val);
    }

    return val;
}

71
uint64_t cpu_to_dump64(DumpState *s, uint64_t val)
72
{
73
    if (s->dump_info.d_endian == ELFDATA2LSB) {
74 75 76 77 78 79 80 81 82 83
        val = cpu_to_le64(val);
    } else {
        val = cpu_to_be64(val);
    }

    return val;
}

static int dump_cleanup(DumpState *s)
{
84
    guest_phys_blocks_free(&s->guest_phys_blocks);
85
    memory_mapping_list_free(&s->list);
86
    close(s->fd);
87 88
    g_free(s->guest_note);
    s->guest_note = NULL;
89
    if (s->resume) {
90 91 92
        if (s->detached) {
            qemu_mutex_lock_iothread();
        }
93
        vm_start();
94 95 96
        if (s->detached) {
            qemu_mutex_unlock_iothread();
        }
97 98
    }

99
    return 0;
100 101
}

102
static int fd_write_vmcore(const void *buf, size_t size, void *opaque)
103 104
{
    DumpState *s = opaque;
105 106 107 108 109
    size_t written_size;

    written_size = qemu_write_full(s->fd, buf, size);
    if (written_size != size) {
        return -1;
110 111 112 113 114
    }

    return 0;
}

115
static void write_elf64_header(DumpState *s, Error **errp)
116 117 118 119 120 121 122 123 124
{
    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;
125 126 127 128 129 130 131
    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);
132 133 134
    if (s->have_section) {
        uint64_t shoff = sizeof(Elf64_Ehdr) + sizeof(Elf64_Phdr) * s->sh_info;

135 136 137
        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);
138 139 140 141
    }

    ret = fd_write_vmcore(&elf_header, sizeof(elf_header), s);
    if (ret < 0) {
142
        error_setg(errp, "dump: failed to write elf header");
143 144 145
    }
}

146
static void write_elf32_header(DumpState *s, Error **errp)
147 148 149 150 151 152 153
{
    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;
154
    elf_header.e_ident[EI_DATA] = s->dump_info.d_endian;
155
    elf_header.e_ident[EI_VERSION] = EV_CURRENT;
156 157 158 159 160 161 162
    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);
163 164 165
    if (s->have_section) {
        uint32_t shoff = sizeof(Elf32_Ehdr) + sizeof(Elf32_Phdr) * s->sh_info;

166 167 168
        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);
169 170 171 172
    }

    ret = fd_write_vmcore(&elf_header, sizeof(elf_header), s);
    if (ret < 0) {
173
        error_setg(errp, "dump: failed to write elf header");
174 175 176
    }
}

177 178 179
static void write_elf64_load(DumpState *s, MemoryMapping *memory_mapping,
                             int phdr_index, hwaddr offset,
                             hwaddr filesz, Error **errp)
180 181 182 183 184
{
    Elf64_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf64_Phdr));
185 186 187 188 189 190
    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);
191

192 193
    assert(memory_mapping->length >= filesz);

194 195
    ret = fd_write_vmcore(&phdr, sizeof(Elf64_Phdr), s);
    if (ret < 0) {
196
        error_setg(errp, "dump: failed to write program header table");
197 198 199
    }
}

200 201 202
static void write_elf32_load(DumpState *s, MemoryMapping *memory_mapping,
                             int phdr_index, hwaddr offset,
                             hwaddr filesz, Error **errp)
203 204 205 206 207
{
    Elf32_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf32_Phdr));
208 209 210 211 212 213
    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);
214

215 216
    assert(memory_mapping->length >= filesz);

217 218
    ret = fd_write_vmcore(&phdr, sizeof(Elf32_Phdr), s);
    if (ret < 0) {
219
        error_setg(errp, "dump: failed to write program header table");
220 221 222
    }
}

223
static void write_elf64_note(DumpState *s, Error **errp)
224 225
{
    Elf64_Phdr phdr;
226
    hwaddr begin = s->memory_offset - s->note_size;
227 228 229
    int ret;

    memset(&phdr, 0, sizeof(Elf64_Phdr));
230 231
    phdr.p_type = cpu_to_dump32(s, PT_NOTE);
    phdr.p_offset = cpu_to_dump64(s, begin);
232
    phdr.p_paddr = 0;
233 234
    phdr.p_filesz = cpu_to_dump64(s, s->note_size);
    phdr.p_memsz = cpu_to_dump64(s, s->note_size);
235 236 237 238
    phdr.p_vaddr = 0;

    ret = fd_write_vmcore(&phdr, sizeof(Elf64_Phdr), s);
    if (ret < 0) {
239
        error_setg(errp, "dump: failed to write program header table");
240 241 242
    }
}

243 244 245 246 247
static inline int cpu_index(CPUState *cpu)
{
    return cpu->cpu_index + 1;
}

248 249 250 251 252 253 254 255 256 257 258 259 260
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");
        }
    }
}

261 262
static void write_elf64_notes(WriteCoreDumpFunction f, DumpState *s,
                              Error **errp)
263
{
264
    CPUState *cpu;
265 266 267
    int ret;
    int id;

268
    CPU_FOREACH(cpu) {
269
        id = cpu_index(cpu);
270
        ret = cpu_write_elf64_note(f, cpu, id, s);
271
        if (ret < 0) {
272
            error_setg(errp, "dump: failed to write elf notes");
273
            return;
274 275 276
        }
    }

277
    CPU_FOREACH(cpu) {
278
        ret = cpu_write_elf64_qemunote(f, cpu, s);
279
        if (ret < 0) {
280
            error_setg(errp, "dump: failed to write CPU status");
281
            return;
282 283
        }
    }
284 285

    write_guest_note(f, s, errp);
286 287
}

288
static void write_elf32_note(DumpState *s, Error **errp)
289
{
290
    hwaddr begin = s->memory_offset - s->note_size;
291 292 293 294
    Elf32_Phdr phdr;
    int ret;

    memset(&phdr, 0, sizeof(Elf32_Phdr));
295 296
    phdr.p_type = cpu_to_dump32(s, PT_NOTE);
    phdr.p_offset = cpu_to_dump32(s, begin);
297
    phdr.p_paddr = 0;
298 299
    phdr.p_filesz = cpu_to_dump32(s, s->note_size);
    phdr.p_memsz = cpu_to_dump32(s, s->note_size);
300 301 302 303
    phdr.p_vaddr = 0;

    ret = fd_write_vmcore(&phdr, sizeof(Elf32_Phdr), s);
    if (ret < 0) {
304
        error_setg(errp, "dump: failed to write program header table");
305 306 307
    }
}

308 309
static void write_elf32_notes(WriteCoreDumpFunction f, DumpState *s,
                              Error **errp)
310
{
311
    CPUState *cpu;
312 313 314
    int ret;
    int id;

315
    CPU_FOREACH(cpu) {
316
        id = cpu_index(cpu);
317
        ret = cpu_write_elf32_note(f, cpu, id, s);
318
        if (ret < 0) {
319
            error_setg(errp, "dump: failed to write elf notes");
320
            return;
321 322 323
        }
    }

324
    CPU_FOREACH(cpu) {
325
        ret = cpu_write_elf32_qemunote(f, cpu, s);
326
        if (ret < 0) {
327
            error_setg(errp, "dump: failed to write CPU status");
328
            return;
329 330
        }
    }
331 332

    write_guest_note(f, s, errp);
333 334
}

335
static void write_elf_section(DumpState *s, int type, Error **errp)
336 337 338 339 340 341 342 343 344 345
{
    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);
346
        shdr32.sh_info = cpu_to_dump32(s, s->sh_info);
347 348 349 350
        shdr = &shdr32;
    } else {
        shdr_size = sizeof(Elf64_Shdr);
        memset(&shdr64, 0, shdr_size);
351
        shdr64.sh_info = cpu_to_dump32(s, s->sh_info);
352 353 354 355 356
        shdr = &shdr64;
    }

    ret = fd_write_vmcore(&shdr, shdr_size, s);
    if (ret < 0) {
357
        error_setg(errp, "dump: failed to write section header table");
358 359 360
    }
}

361
static void write_data(DumpState *s, void *buf, int length, Error **errp)
362 363 364 365 366
{
    int ret;

    ret = fd_write_vmcore(buf, length, s);
    if (ret < 0) {
367
        error_setg(errp, "dump: failed to save memory");
368 369
    } else {
        s->written_size += length;
370 371 372
    }
}

373 374 375
/* 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)
376 377
{
    int64_t i;
378
    Error *local_err = NULL;
379

380 381 382
    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);
383 384 385
        if (local_err) {
            error_propagate(errp, local_err);
            return;
386 387 388
        }
    }

389 390 391
    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);
392 393 394
        if (local_err) {
            error_propagate(errp, local_err);
            return;
395 396 397 398
        }
    }
}

399 400 401 402 403 404
/* 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)
405
{
406
    GuestPhysBlock *block;
407
    hwaddr offset = s->memory_offset;
408 409
    int64_t size_in_block, start;

410 411 412 413
    /* When the memory is not stored into vmcore, offset will be -1 */
    *p_offset = -1;
    *p_filesz = 0;

414 415
    if (s->has_filter) {
        if (phys_addr < s->begin || phys_addr >= s->begin + s->length) {
416
            return;
417 418 419
        }
    }

420
    QTAILQ_FOREACH(block, &s->guest_phys_blocks.head, next) {
421
        if (s->has_filter) {
422 423
            if (block->target_start >= s->begin + s->length ||
                block->target_end <= s->begin) {
424 425 426 427
                /* This block is out of the range */
                continue;
            }

428 429
            if (s->begin <= block->target_start) {
                start = block->target_start;
430 431 432 433
            } else {
                start = s->begin;
            }

434 435 436
            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);
437 438
            }
        } else {
439 440
            start = block->target_start;
            size_in_block = block->target_end - block->target_start;
441 442 443
        }

        if (phys_addr >= start && phys_addr < start + size_in_block) {
444 445 446
            *p_offset = phys_addr - start + offset;

            /* The offset range mapped from the vmcore file must not spill over
447
             * the GuestPhysBlock, clamp it. The rest of the mapping will be
448 449 450 451 452 453 454
             * 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;
455 456 457 458 459 460
        }

        offset += size_in_block;
    }
}

461
static void write_elf_loads(DumpState *s, Error **errp)
462
{
463
    hwaddr offset, filesz;
464 465 466
    MemoryMapping *memory_mapping;
    uint32_t phdr_index = 1;
    uint32_t max_index;
467
    Error *local_err = NULL;
468 469 470 471 472 473 474 475

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

    QTAILQ_FOREACH(memory_mapping, &s->list.head, next) {
476 477 478
        get_offset_range(memory_mapping->phys_addr,
                         memory_mapping->length,
                         s, &offset, &filesz);
479
        if (s->dump_info.d_class == ELFCLASS64) {
480 481
            write_elf64_load(s, memory_mapping, phdr_index++, offset,
                             filesz, &local_err);
482
        } else {
483 484
            write_elf32_load(s, memory_mapping, phdr_index++, offset,
                             filesz, &local_err);
485 486
        }

487 488 489
        if (local_err) {
            error_propagate(errp, local_err);
            return;
490 491 492 493 494 495 496 497 498
        }

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

/* write elf header, PT_NOTE and elf note to vmcore. */
499
static void dump_begin(DumpState *s, Error **errp)
500
{
501
    Error *local_err = NULL;
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528

    /*
     * 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) {
529
        write_elf64_header(s, &local_err);
530
    } else {
531
        write_elf32_header(s, &local_err);
532
    }
533 534 535
    if (local_err) {
        error_propagate(errp, local_err);
        return;
536 537 538 539
    }

    if (s->dump_info.d_class == ELFCLASS64) {
        /* write PT_NOTE to vmcore */
540 541 542 543
        write_elf64_note(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
544 545 546
        }

        /* write all PT_LOAD to vmcore */
547 548 549 550
        write_elf_loads(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
551 552 553 554
        }

        /* write section to vmcore */
        if (s->have_section) {
555 556 557 558
            write_elf_section(s, 1, &local_err);
            if (local_err) {
                error_propagate(errp, local_err);
                return;
559 560 561 562
            }
        }

        /* write notes to vmcore */
563 564 565 566
        write_elf64_notes(fd_write_vmcore, s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
567 568 569
        }
    } else {
        /* write PT_NOTE to vmcore */
570 571 572 573
        write_elf32_note(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
574 575 576
        }

        /* write all PT_LOAD to vmcore */
577 578 579 580
        write_elf_loads(s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
581 582 583 584
        }

        /* write section to vmcore */
        if (s->have_section) {
585 586 587 588
            write_elf_section(s, 0, &local_err);
            if (local_err) {
                error_propagate(errp, local_err);
                return;
589 590 591 592
            }
        }

        /* write notes to vmcore */
593 594 595 596
        write_elf32_notes(fd_write_vmcore, s, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
597 598 599 600
        }
    }
}

601
static int get_next_block(DumpState *s, GuestPhysBlock *block)
602 603
{
    while (1) {
604
        block = QTAILQ_NEXT(block, next);
605 606 607 608 609 610
        if (!block) {
            /* no more block */
            return 1;
        }

        s->start = 0;
611
        s->next_block = block;
612
        if (s->has_filter) {
613 614
            if (block->target_start >= s->begin + s->length ||
                block->target_end <= s->begin) {
615 616 617 618
                /* This block is out of the range */
                continue;
            }

619 620
            if (s->begin > block->target_start) {
                s->start = s->begin - block->target_start;
621 622 623 624 625 626 627 628
            }
        }

        return 0;
    }
}

/* write all memory to vmcore */
629
static void dump_iterate(DumpState *s, Error **errp)
630
{
631
    GuestPhysBlock *block;
632
    int64_t size;
633
    Error *local_err = NULL;
634

635
    do {
636
        block = s->next_block;
637

638
        size = block->target_end - block->target_start;
639 640
        if (s->has_filter) {
            size -= s->start;
641 642
            if (s->begin + s->length < block->target_end) {
                size -= block->target_end - (s->begin + s->length);
643 644
            }
        }
645 646 647 648
        write_memory(s, block, s->start, size, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return;
649 650
        }

651
    } while (!get_next_block(s, block));
652 653
}

654
static void create_vmcore(DumpState *s, Error **errp)
655
{
656
    Error *local_err = NULL;
657

658 659 660 661
    dump_begin(s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
662 663
    }

664
    dump_iterate(s, errp);
665 666
}

667 668
static int write_start_flat_header(int fd)
{
669
    MakedumpfileHeader *mh;
670 671
    int ret = 0;

672 673
    QEMU_BUILD_BUG_ON(sizeof *mh > MAX_SIZE_MDF_HEADER);
    mh = g_malloc0(MAX_SIZE_MDF_HEADER);
674

675 676
    memcpy(mh->signature, MAKEDUMPFILE_SIGNATURE,
           MIN(sizeof mh->signature, sizeof MAKEDUMPFILE_SIGNATURE));
677

678 679
    mh->type = cpu_to_be64(TYPE_FLAT_HEADER);
    mh->version = cpu_to_be64(VERSION_FLAT_HEADER);
680 681

    size_t written_size;
682
    written_size = qemu_write_full(fd, mh, MAX_SIZE_MDF_HEADER);
683 684 685 686
    if (written_size != MAX_SIZE_MDF_HEADER) {
        ret = -1;
    }

687
    g_free(mh);
688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706
    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;
}

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727
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;
}

728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
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;
}

744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781
/*
 * 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;
    }
}

782 783 784 785 786 787 788 789 790
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);

791
    return name_size == len && memcmp(note + head_size, name, len) == 0;
792 793
}

794
/* write common header, sub header and elf note to vmcore */
795
static void create_header32(DumpState *s, Error **errp)
796 797 798 799 800 801 802 803 804
{
    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;
805
    Error *local_err = NULL;
806 807 808 809 810 811

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

    strncpy(dh->signature, KDUMP_SIGNATURE, strlen(KDUMP_SIGNATURE));
812
    dh->header_version = cpu_to_dump32(s, 6);
813
    block_size = s->dump_info.page_size;
814
    dh->block_size = cpu_to_dump32(s, block_size);
815 816
    sub_hdr_size = sizeof(struct KdumpSubHeader32) + s->note_size;
    sub_hdr_size = DIV_ROUND_UP(sub_hdr_size, block_size);
817
    dh->sub_hdr_size = cpu_to_dump32(s, sub_hdr_size);
818
    /* dh->max_mapnr may be truncated, full 64bit is in kh.max_mapnr_64 */
819 820
    dh->max_mapnr = cpu_to_dump32(s, MIN(s->max_mapnr, UINT_MAX));
    dh->nr_cpus = cpu_to_dump32(s, s->nr_cpus);
821
    bitmap_blocks = DIV_ROUND_UP(s->len_dump_bitmap, block_size) * 2;
822
    dh->bitmap_blocks = cpu_to_dump32(s, bitmap_blocks);
823
    strncpy(dh->utsname.machine, ELF_MACHINE_UNAME, sizeof(dh->utsname.machine));
824 825 826 827 828 829 830 831 832 833 834 835 836 837

    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
838
    dh->status = cpu_to_dump32(s, status);
839 840

    if (write_buffer(s->fd, 0, dh, size) < 0) {
841
        error_setg(errp, "dump: failed to write disk dump header");
842 843 844 845 846 847 848 849
        goto out;
    }

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

    /* 64bit max_mapnr_64 */
850
    kh->max_mapnr_64 = cpu_to_dump64(s, s->max_mapnr);
851
    kh->phys_base = cpu_to_dump32(s, s->dump_info.phys_base);
852
    kh->dump_level = cpu_to_dump32(s, DUMP_LEVEL);
853 854

    offset_note = DISKDUMP_HEADER_BLOCKS * block_size + size;
855 856 857 858 859 860 861 862 863 864 865 866
    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);
    }

867 868
    kh->offset_note = cpu_to_dump64(s, offset_note);
    kh->note_size = cpu_to_dump32(s, s->note_size);
869 870 871

    if (write_buffer(s->fd, DISKDUMP_HEADER_BLOCKS *
                     block_size, kh, size) < 0) {
872
        error_setg(errp, "dump: failed to write kdump sub header");
873 874 875 876 877 878 879 880
        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 */
881 882 883
    write_elf32_notes(buf_write_note, s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
884 885 886 887
        goto out;
    }
    if (write_buffer(s->fd, offset_note, s->note_buf,
                     s->note_size) < 0) {
888
        error_setg(errp, "dump: failed to write notes");
889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906
        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 */
907
static void create_header64(DumpState *s, Error **errp)
908 909 910 911 912 913 914 915 916
{
    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;
917
    Error *local_err = NULL;
918 919 920 921 922 923

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

    strncpy(dh->signature, KDUMP_SIGNATURE, strlen(KDUMP_SIGNATURE));
924
    dh->header_version = cpu_to_dump32(s, 6);
925
    block_size = s->dump_info.page_size;
926
    dh->block_size = cpu_to_dump32(s, block_size);
927 928
    sub_hdr_size = sizeof(struct KdumpSubHeader64) + s->note_size;
    sub_hdr_size = DIV_ROUND_UP(sub_hdr_size, block_size);
929
    dh->sub_hdr_size = cpu_to_dump32(s, sub_hdr_size);
930
    /* dh->max_mapnr may be truncated, full 64bit is in kh.max_mapnr_64 */
931 932
    dh->max_mapnr = cpu_to_dump32(s, MIN(s->max_mapnr, UINT_MAX));
    dh->nr_cpus = cpu_to_dump32(s, s->nr_cpus);
933
    bitmap_blocks = DIV_ROUND_UP(s->len_dump_bitmap, block_size) * 2;
934
    dh->bitmap_blocks = cpu_to_dump32(s, bitmap_blocks);
935
    strncpy(dh->utsname.machine, ELF_MACHINE_UNAME, sizeof(dh->utsname.machine));
936 937 938 939 940 941 942 943 944 945 946 947 948 949

    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
950
    dh->status = cpu_to_dump32(s, status);
951 952

    if (write_buffer(s->fd, 0, dh, size) < 0) {
953
        error_setg(errp, "dump: failed to write disk dump header");
954 955 956 957 958 959 960 961
        goto out;
    }

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

    /* 64bit max_mapnr_64 */
962
    kh->max_mapnr_64 = cpu_to_dump64(s, s->max_mapnr);
963
    kh->phys_base = cpu_to_dump64(s, s->dump_info.phys_base);
964
    kh->dump_level = cpu_to_dump32(s, DUMP_LEVEL);
965 966

    offset_note = DISKDUMP_HEADER_BLOCKS * block_size + size;
967 968 969 970 971 972 973 974 975 976 977 978
    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);
    }

979 980
    kh->offset_note = cpu_to_dump64(s, offset_note);
    kh->note_size = cpu_to_dump64(s, s->note_size);
981 982 983

    if (write_buffer(s->fd, DISKDUMP_HEADER_BLOCKS *
                     block_size, kh, size) < 0) {
984
        error_setg(errp, "dump: failed to write kdump sub header");
985 986 987 988 989 990 991 992
        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 */
993 994 995
    write_elf64_notes(buf_write_note, s, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
996 997 998 999 1000
        goto out;
    }

    if (write_buffer(s->fd, offset_note, s->note_buf,
                     s->note_size) < 0) {
1001
        error_setg(errp, "dump: failed to write notes");
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
        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);
}

1019
static void write_dump_header(DumpState *s, Error **errp)
1020
{
1021 1022
     Error *local_err = NULL;

1023
    if (s->dump_info.d_class == ELFCLASS32) {
1024
        create_header32(s, &local_err);
1025
    } else {
1026 1027
        create_header64(s, &local_err);
    }
1028
    error_propagate(errp, local_err);
1029 1030
}

1031 1032 1033 1034 1035
static size_t dump_bitmap_get_bufsize(DumpState *s)
{
    return s->dump_info.page_size;
}

1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048
/*
 * 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;
1049 1050
    size_t bitmap_bufsize = dump_bitmap_get_bufsize(s);
    size_t bits_per_buf = bitmap_bufsize * CHAR_BIT;
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060

    /* 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.
     */
1061 1062
    old_offset = bitmap_bufsize * (last_pfn / bits_per_buf);
    new_offset = bitmap_bufsize * (pfn / bits_per_buf);
1063 1064 1065 1066 1067

    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,
1068
                         bitmap_bufsize) < 0) {
1069 1070 1071 1072 1073 1074 1075
            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,
1076
                         bitmap_bufsize) < 0) {
1077 1078 1079
            return -1;
        }

1080 1081
        memset(buf, 0, bitmap_bufsize);
        old_offset += bitmap_bufsize;
1082 1083 1084
    }

    /* get the exact place of the bit in the buf, and set it */
1085 1086
    byte = (pfn % bits_per_buf) / CHAR_BIT;
    bit = (pfn % bits_per_buf) % CHAR_BIT;
1087 1088 1089 1090 1091 1092 1093 1094 1095
    if (value) {
        buf[byte] |= 1u << bit;
    } else {
        buf[byte] &= ~(1u << bit);
    }

    return 0;
}

1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
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;
}

1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
/*
 * 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;
1120
    hwaddr addr, target_page_mask = ~((hwaddr)s->dump_info.page_size - 1);
1121 1122 1123 1124 1125 1126
    uint8_t *buf;

    /* block == NULL means the start of the iteration */
    if (!block) {
        block = QTAILQ_FIRST(&s->guest_phys_blocks.head);
        *blockptr = block;
1127 1128 1129
        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);
1130 1131 1132 1133 1134 1135 1136
        if (bufptr) {
            *bufptr = block->host_addr;
        }
        return true;
    }

    *pfnptr = *pfnptr + 1;
1137
    addr = dump_pfn_to_paddr(s, *pfnptr);
1138 1139

    if ((addr >= block->target_start) &&
1140
        (addr + s->dump_info.page_size <= block->target_end)) {
1141 1142 1143 1144 1145 1146 1147 1148
        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;
        }
1149 1150 1151
        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);
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
        buf = block->host_addr;
    }

    if (bufptr) {
        *bufptr = buf;
    }

    return true;
}

1162
static void write_dump_bitmap(DumpState *s, Error **errp)
1163 1164 1165 1166 1167 1168
{
    int ret = 0;
    uint64_t last_pfn, pfn;
    void *dump_bitmap_buf;
    size_t num_dumpable;
    GuestPhysBlock *block_iter = NULL;
1169 1170
    size_t bitmap_bufsize = dump_bitmap_get_bufsize(s);
    size_t bits_per_buf = bitmap_bufsize * CHAR_BIT;
1171 1172

    /* dump_bitmap_buf is used to store dump_bitmap temporarily */
1173
    dump_bitmap_buf = g_malloc0(bitmap_bufsize);
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184

    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) {
1185
            error_setg(errp, "dump: failed to set dump_bitmap");
1186 1187 1188 1189 1190 1191 1192 1193 1194
            goto out;
        }

        last_pfn = pfn;
        num_dumpable++;
    }

    /*
     * set_dump_bitmap will always leave the recently set bit un-sync. Here we
1195 1196
     * 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.
1197 1198
     */
    if (num_dumpable > 0) {
1199
        ret = set_dump_bitmap(last_pfn, last_pfn + bits_per_buf, false,
1200 1201
                              dump_bitmap_buf, s);
        if (ret < 0) {
1202
            error_setg(errp, "dump: failed to sync dump_bitmap");
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
            goto out;
        }
    }

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

out:
    g_free(dump_bitmap_buf);
}

1214 1215 1216 1217 1218
static void prepare_data_cache(DataCache *data_cache, DumpState *s,
                               off_t offset)
{
    data_cache->fd = s->fd;
    data_cache->data_size = 0;
1219 1220
    data_cache->buf_size = 4 * dump_bitmap_get_bufsize(s);
    data_cache->buf = g_malloc0(data_cache->buf_size);
1221 1222 1223 1224 1225 1226 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
    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);
}

1261 1262
static size_t get_len_buf_out(size_t page_size, uint32_t flag_compress)
{
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
    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;
1274 1275

#ifdef CONFIG_SNAPPY
1276 1277
    case DUMP_DH_COMPRESSED_SNAPPY:
        return snappy_max_compressed_length(page_size);
1278
#endif
1279 1280
    }
    return 0;
1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
}

/*
 * 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);
}

1291
static void write_dump_pages(DumpState *s, Error **errp)
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313
{
    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 */
1314
    len_buf_out = get_len_buf_out(s->dump_info.page_size, s->flag_compress);
1315
    assert(len_buf_out != 0);
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326

#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
     */
1327
    pd_zero.size = cpu_to_dump32(s, s->dump_info.page_size);
1328 1329 1330
    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);
1331 1332
    buf = g_malloc0(s->dump_info.page_size);
    ret = write_cache(&page_data, buf, s->dump_info.page_size, false);
1333 1334
    g_free(buf);
    if (ret < 0) {
1335
        error_setg(errp, "dump: failed to write page data (zero page)");
1336 1337 1338
        goto out;
    }

1339
    offset_data += s->dump_info.page_size;
1340 1341 1342 1343 1344 1345 1346

    /*
     * 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 */
1347
        if (is_zero_page(buf, s->dump_info.page_size)) {
1348 1349 1350
            ret = write_cache(&page_desc, &pd_zero, sizeof(PageDescriptor),
                              false);
            if (ret < 0) {
1351
                error_setg(errp, "dump: failed to write page desc");
1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
                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) &&
1368
                    (compress2(buf_out, (uLongf *)&size_out, buf,
1369 1370
                               s->dump_info.page_size, Z_BEST_SPEED) == Z_OK) &&
                    (size_out < s->dump_info.page_size)) {
1371 1372
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_ZLIB);
                pd.size  = cpu_to_dump32(s, size_out);
1373 1374 1375

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1376
                    error_setg(errp, "dump: failed to write page data");
1377 1378 1379 1380
                    goto out;
                }
#ifdef CONFIG_LZO
            } else if ((s->flag_compress & DUMP_DH_COMPRESSED_LZO) &&
1381
                    (lzo1x_1_compress(buf, s->dump_info.page_size, buf_out,
1382
                    (lzo_uint *)&size_out, wrkmem) == LZO_E_OK) &&
1383
                    (size_out < s->dump_info.page_size)) {
1384 1385
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_LZO);
                pd.size  = cpu_to_dump32(s, size_out);
1386 1387 1388

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1389
                    error_setg(errp, "dump: failed to write page data");
1390 1391 1392 1393 1394
                    goto out;
                }
#endif
#ifdef CONFIG_SNAPPY
            } else if ((s->flag_compress & DUMP_DH_COMPRESSED_SNAPPY) &&
1395
                    (snappy_compress((char *)buf, s->dump_info.page_size,
1396
                    (char *)buf_out, &size_out) == SNAPPY_OK) &&
1397
                    (size_out < s->dump_info.page_size)) {
1398 1399
                pd.flags = cpu_to_dump32(s, DUMP_DH_COMPRESSED_SNAPPY);
                pd.size  = cpu_to_dump32(s, size_out);
1400 1401 1402

                ret = write_cache(&page_data, buf_out, size_out, false);
                if (ret < 0) {
1403
                    error_setg(errp, "dump: failed to write page data");
1404 1405 1406 1407 1408 1409
                    goto out;
                }
#endif
            } else {
                /*
                 * fall back to save in plaintext, size_out should be
1410
                 * assigned the target's page size
1411
                 */
1412
                pd.flags = cpu_to_dump32(s, 0);
1413
                size_out = s->dump_info.page_size;
1414
                pd.size = cpu_to_dump32(s, size_out);
1415

1416 1417
                ret = write_cache(&page_data, buf,
                                  s->dump_info.page_size, false);
1418
                if (ret < 0) {
1419
                    error_setg(errp, "dump: failed to write page data");
1420 1421 1422 1423 1424
                    goto out;
                }
            }

            /* get and write page desc here */
1425 1426
            pd.page_flags = cpu_to_dump64(s, 0);
            pd.offset = cpu_to_dump64(s, offset_data);
1427 1428 1429 1430
            offset_data += size_out;

            ret = write_cache(&page_desc, &pd, sizeof(PageDescriptor), false);
            if (ret < 0) {
1431
                error_setg(errp, "dump: failed to write page desc");
1432 1433 1434
                goto out;
            }
        }
Peter Xu's avatar