Quellcodebibliothek Statistik Leitseite products/Sources/formale Sprachen/C/Linux/drivers/md/dm-vdo/   (Open Source Betriebssystem Version 6.17.9©)  Datei vom 24.10.2025 mit Größe 45 kB image not shown  

Quelle  encodings.c   Sprache: C

 
// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright 2023 Red Hat
 */


#include "encodings.h"

#include <linux/log2.h>

#include "logger.h"
#include "memory-alloc.h"
#include "permassert.h"

#include "constants.h"
#include "status-codes.h"
#include "types.h"

/** The maximum logical space is 4 petabytes, which is 1 terablock. */
static const block_count_t MAXIMUM_VDO_LOGICAL_BLOCKS = 1024ULL * 1024 * 1024 * 1024;

/** The maximum physical space is 256 terabytes, which is 64 gigablocks. */
static const block_count_t MAXIMUM_VDO_PHYSICAL_BLOCKS = 1024ULL * 1024 * 1024 * 64;

struct geometry_block {
 char magic_number[VDO_GEOMETRY_MAGIC_NUMBER_SIZE];
 struct packed_header header;
 u32 checksum;
} __packed;

static const struct header GEOMETRY_BLOCK_HEADER_5_0 = {
 .id = VDO_GEOMETRY_BLOCK,
 .version = {
  .major_version = 5,
  .minor_version = 0,
 },
 /*
 * Note: this size isn't just the payload size following the header, like it is everywhere
 * else in VDO.
 */

 .size = sizeof(struct geometry_block) + sizeof(struct volume_geometry),
};

static const struct header GEOMETRY_BLOCK_HEADER_4_0 = {
 .id = VDO_GEOMETRY_BLOCK,
 .version = {
  .major_version = 4,
  .minor_version = 0,
 },
 /*
 * Note: this size isn't just the payload size following the header, like it is everywhere
 * else in VDO.
 */

 .size = sizeof(struct geometry_block) + sizeof(struct volume_geometry_4_0),
};

const u8 VDO_GEOMETRY_MAGIC_NUMBER[VDO_GEOMETRY_MAGIC_NUMBER_SIZE + 1] = "dmvdo001";

#define PAGE_HEADER_4_1_SIZE (8 + 8 + 8 + 1 + 1 + 1 + 1)

static const struct version_number BLOCK_MAP_4_1 = {
 .major_version = 4,
 .minor_version = 1,
};

const struct header VDO_BLOCK_MAP_HEADER_2_0 = {
 .id = VDO_BLOCK_MAP,
 .version = {
  .major_version = 2,
  .minor_version = 0,
 },
 .size = sizeof(struct block_map_state_2_0),
};

const struct header VDO_RECOVERY_JOURNAL_HEADER_7_0 = {
 .id = VDO_RECOVERY_JOURNAL,
 .version = {
   .major_version = 7,
   .minor_version = 0,
  },
 .size = sizeof(struct recovery_journal_state_7_0),
};

const struct header VDO_SLAB_DEPOT_HEADER_2_0 = {
 .id = VDO_SLAB_DEPOT,
 .version = {
  .major_version = 2,
  .minor_version = 0,
 },
 .size = sizeof(struct slab_depot_state_2_0),
};

static const struct header VDO_LAYOUT_HEADER_3_0 = {
 .id = VDO_LAYOUT,
 .version = {
  .major_version = 3,
  .minor_version = 0,
 },
 .size = sizeof(struct layout_3_0) + (sizeof(struct partition_3_0) * VDO_PARTITION_COUNT),
};

static const enum partition_id REQUIRED_PARTITIONS[] = {
 VDO_BLOCK_MAP_PARTITION,
 VDO_SLAB_DEPOT_PARTITION,
 VDO_RECOVERY_JOURNAL_PARTITION,
 VDO_SLAB_SUMMARY_PARTITION,
};

/*
 * The current version for the data encoded in the super block. This must be changed any time there
 * is a change to encoding of the component data of any VDO component.
 */

static const struct version_number VDO_COMPONENT_DATA_41_0 = {
 .major_version = 41,
 .minor_version = 0,
};

const struct version_number VDO_VOLUME_VERSION_67_0 = {
 .major_version = 67,
 .minor_version = 0,
};

static const struct header SUPER_BLOCK_HEADER_12_0 = {
 .id = VDO_SUPER_BLOCK,
 .version = {
   .major_version = 12,
   .minor_version = 0,
  },

 /* This is the minimum size, if the super block contains no components. */
 .size = VDO_SUPER_BLOCK_FIXED_SIZE - VDO_ENCODED_HEADER_SIZE,
};

/**
 * validate_version() - Check whether a version matches an expected version.
 * @expected_version: The expected version.
 * @actual_version: The version being validated.
 * @component_name: The name of the component or the calling function (for error logging).
 *
 * Logs an error describing a mismatch.
 *
 * Return: VDO_SUCCESS             if the versions are the same,
 *         VDO_UNSUPPORTED_VERSION if the versions don't match.
 */

static int __must_check validate_version(struct version_number expected_version,
      struct version_number actual_version,
      const char *component_name)
{
 if (!vdo_are_same_version(expected_version, actual_version)) {
  return vdo_log_error_strerror(VDO_UNSUPPORTED_VERSION,
           "%s version mismatch, expected %d.%d, got %d.%d",
           component_name,
           expected_version.major_version,
           expected_version.minor_version,
           actual_version.major_version,
           actual_version.minor_version);
 }

 return VDO_SUCCESS;
}

/**
 * vdo_validate_header() - Check whether a header matches expectations.
 * @expected_header: The expected header.
 * @actual_header: The header being validated.
 * @exact_size: If true, the size fields of the two headers must be the same, otherwise it is
 *              required that actual_header.size >= expected_header.size.
 * @name: The name of the component or the calling function (for error logging).
 *
 * Logs an error describing the first mismatch found.
 *
 * Return: VDO_SUCCESS             if the header meets expectations,
 *         VDO_INCORRECT_COMPONENT if the component ids don't match,
 *         VDO_UNSUPPORTED_VERSION if the versions or sizes don't match.
 */

int vdo_validate_header(const struct header *expected_header,
   const struct header *actual_header, bool exact_size,
   const char *name)
{
 int result;

 if (expected_header->id != actual_header->id) {
  return vdo_log_error_strerror(VDO_INCORRECT_COMPONENT,
           "%s ID mismatch, expected %d, got %d",
           name, expected_header->id,
           actual_header->id);
 }

 result = validate_version(expected_header->version, actual_header->version,
      name);
 if (result != VDO_SUCCESS)
  return result;

 if ((expected_header->size > actual_header->size) ||
     (exact_size && (expected_header->size < actual_header->size))) {
  return vdo_log_error_strerror(VDO_UNSUPPORTED_VERSION,
           "%s size mismatch, expected %zu, got %zu",
           name, expected_header->size,
           actual_header->size);
 }

 return VDO_SUCCESS;
}

static void encode_version_number(u8 *buffer, size_t *offset,
      struct version_number version)
{
 struct packed_version_number packed = vdo_pack_version_number(version);

 memcpy(buffer + *offset, &packed, sizeof(packed));
 *offset += sizeof(packed);
}

void vdo_encode_header(u8 *buffer, size_t *offset, const struct header *header)
{
 struct packed_header packed = vdo_pack_header(header);

 memcpy(buffer + *offset, &packed, sizeof(packed));
 *offset += sizeof(packed);
}

static void decode_version_number(u8 *buffer, size_t *offset,
      struct version_number *version)
{
 struct packed_version_number packed;

 memcpy(&packed, buffer + *offset, sizeof(packed));
 *offset += sizeof(packed);
 *version = vdo_unpack_version_number(packed);
}

void vdo_decode_header(u8 *buffer, size_t *offset, struct header *header)
{
 struct packed_header packed;

 memcpy(&packed, buffer + *offset, sizeof(packed));
 *offset += sizeof(packed);

 *header = vdo_unpack_header(&packed);
}

/**
 * decode_volume_geometry() - Decode the on-disk representation of a volume geometry from a buffer.
 * @buffer: A buffer to decode from.
 * @offset: The offset in the buffer at which to decode.
 * @geometry: The structure to receive the decoded fields.
 * @version: The geometry block version to decode.
 */

static void decode_volume_geometry(u8 *buffer, size_t *offset,
       struct volume_geometry *geometry, u32 version)
{
 u32 unused, mem;
 enum volume_region_id id;
 nonce_t nonce;
 block_count_t bio_offset = 0;
 bool sparse;

 /* This is for backwards compatibility. */
 decode_u32_le(buffer, offset, &unused);
 geometry->unused = unused;

 decode_u64_le(buffer, offset, &nonce);
 geometry->nonce = nonce;

 memcpy((unsigned char *) &geometry->uuid, buffer + *offset, sizeof(uuid_t));
 *offset += sizeof(uuid_t);

 if (version > 4)
  decode_u64_le(buffer, offset, &bio_offset);
 geometry->bio_offset = bio_offset;

 for (id = 0; id < VDO_VOLUME_REGION_COUNT; id++) {
  physical_block_number_t start_block;
  enum volume_region_id saved_id;

  decode_u32_le(buffer, offset, &saved_id);
  decode_u64_le(buffer, offset, &start_block);

  geometry->regions[id] = (struct volume_region) {
   .id = saved_id,
   .start_block = start_block,
  };
 }

 decode_u32_le(buffer, offset, &mem);
 *offset += sizeof(u32);
 sparse = buffer[(*offset)++];

 geometry->index_config = (struct index_config) {
  .mem = mem,
  .sparse = sparse,
 };
}

/**
 * vdo_parse_geometry_block() - Decode and validate an encoded geometry block.
 * @block: The encoded geometry block.
 * @geometry: The structure to receive the decoded fields.
 */

int __must_check vdo_parse_geometry_block(u8 *block, struct volume_geometry *geometry)
{
 u32 checksum, saved_checksum;
 struct header header;
 size_t offset = 0;
 int result;

 if (memcmp(block, VDO_GEOMETRY_MAGIC_NUMBER, VDO_GEOMETRY_MAGIC_NUMBER_SIZE) != 0)
  return VDO_BAD_MAGIC;
 offset += VDO_GEOMETRY_MAGIC_NUMBER_SIZE;

 vdo_decode_header(block, &offset, &header);
 if (header.version.major_version <= 4) {
  result = vdo_validate_header(&GEOMETRY_BLOCK_HEADER_4_0, &header,
          true, __func__);
 } else {
  result = vdo_validate_header(&GEOMETRY_BLOCK_HEADER_5_0, &header,
          true, __func__);
 }
 if (result != VDO_SUCCESS)
  return result;

 decode_volume_geometry(block, &offset, geometry, header.version.major_version);

 result = VDO_ASSERT(header.size == offset + sizeof(u32),
       "should have decoded up to the geometry checksum");
 if (result != VDO_SUCCESS)
  return result;

 /* Decode and verify the checksum. */
 checksum = vdo_crc32(block, offset);
 decode_u32_le(block, &offset, &saved_checksum);

 return ((checksum == saved_checksum) ? VDO_SUCCESS : VDO_CHECKSUM_MISMATCH);
}

struct block_map_page *vdo_format_block_map_page(void *buffer, nonce_t nonce,
       physical_block_number_t pbn,
       bool initialized)
{
 struct block_map_page *page = buffer;

 memset(buffer, 0, VDO_BLOCK_SIZE);
 page->version = vdo_pack_version_number(BLOCK_MAP_4_1);
 page->header.nonce = __cpu_to_le64(nonce);
 page->header.pbn = __cpu_to_le64(pbn);
 page->header.initialized = initialized;
 return page;
}

enum block_map_page_validity vdo_validate_block_map_page(struct block_map_page *page,
        nonce_t nonce,
        physical_block_number_t pbn)
{
 BUILD_BUG_ON(sizeof(struct block_map_page_header) != PAGE_HEADER_4_1_SIZE);

 if (!vdo_are_same_version(BLOCK_MAP_4_1,
      vdo_unpack_version_number(page->version)) ||
     !page->header.initialized || (nonce != __le64_to_cpu(page->header.nonce)))
  return VDO_BLOCK_MAP_PAGE_INVALID;

 if (pbn != vdo_get_block_map_page_pbn(page))
  return VDO_BLOCK_MAP_PAGE_BAD;

 return VDO_BLOCK_MAP_PAGE_VALID;
}

static int decode_block_map_state_2_0(u8 *buffer, size_t *offset,
          struct block_map_state_2_0 *state)
{
 size_t initial_offset;
 block_count_t flat_page_count, root_count;
 physical_block_number_t flat_page_origin, root_origin;
 struct header header;
 int result;

 vdo_decode_header(buffer, offset, &header);
 result = vdo_validate_header(&VDO_BLOCK_MAP_HEADER_2_0, &header, true, __func__);
 if (result != VDO_SUCCESS)
  return result;

 initial_offset = *offset;

 decode_u64_le(buffer, offset, &flat_page_origin);
 result = VDO_ASSERT(flat_page_origin == VDO_BLOCK_MAP_FLAT_PAGE_ORIGIN,
       "Flat page origin must be %u (recorded as %llu)",
       VDO_BLOCK_MAP_FLAT_PAGE_ORIGIN,
       (unsigned long long) state->flat_page_origin);
 if (result != VDO_SUCCESS)
  return result;

 decode_u64_le(buffer, offset, &flat_page_count);
 result = VDO_ASSERT(flat_page_count == 0,
       "Flat page count must be 0 (recorded as %llu)",
       (unsigned long long) state->flat_page_count);
 if (result != VDO_SUCCESS)
  return result;

 decode_u64_le(buffer, offset, &root_origin);
 decode_u64_le(buffer, offset, &root_count);

 result = VDO_ASSERT(VDO_BLOCK_MAP_HEADER_2_0.size == *offset - initial_offset,
       "decoded block map component size must match header size");
 if (result != VDO_SUCCESS)
  return result;

 *state = (struct block_map_state_2_0) {
  .flat_page_origin = flat_page_origin,
  .flat_page_count = flat_page_count,
  .root_origin = root_origin,
  .root_count = root_count,
 };

 return VDO_SUCCESS;
}

static void encode_block_map_state_2_0(u8 *buffer, size_t *offset,
           struct block_map_state_2_0 state)
{
 size_t initial_offset;

 vdo_encode_header(buffer, offset, &VDO_BLOCK_MAP_HEADER_2_0);

 initial_offset = *offset;
 encode_u64_le(buffer, offset, state.flat_page_origin);
 encode_u64_le(buffer, offset, state.flat_page_count);
 encode_u64_le(buffer, offset, state.root_origin);
 encode_u64_le(buffer, offset, state.root_count);

 VDO_ASSERT_LOG_ONLY(VDO_BLOCK_MAP_HEADER_2_0.size == *offset - initial_offset,
       "encoded block map component size must match header size");
}

/**
 * vdo_compute_new_forest_pages() - Compute the number of pages which must be allocated at each
 *                                  level in order to grow the forest to a new number of entries.
 * @entries: The new number of entries the block map must address.
 *
 * Return: The total number of non-leaf pages required.
 */

block_count_t vdo_compute_new_forest_pages(root_count_t root_count,
        struct boundary *old_sizes,
        block_count_t entries,
        struct boundary *new_sizes)
{
 page_count_t leaf_pages = max(vdo_compute_block_map_page_count(entries), 1U);
 page_count_t level_size = DIV_ROUND_UP(leaf_pages, root_count);
 block_count_t total_pages = 0;
 height_t height;

 for (height = 0; height < VDO_BLOCK_MAP_TREE_HEIGHT; height++) {
  block_count_t new_pages;

  level_size = DIV_ROUND_UP(level_size, VDO_BLOCK_MAP_ENTRIES_PER_PAGE);
  new_sizes->levels[height] = level_size;
  new_pages = level_size;
  if (old_sizes != NULL)
   new_pages -= old_sizes->levels[height];
  total_pages += (new_pages * root_count);
 }

 return total_pages;
}

/**
 * encode_recovery_journal_state_7_0() - Encode the state of a recovery journal.
 *
 * Return: VDO_SUCCESS or an error code.
 */

static void encode_recovery_journal_state_7_0(u8 *buffer, size_t *offset,
           struct recovery_journal_state_7_0 state)
{
 size_t initial_offset;

 vdo_encode_header(buffer, offset, &VDO_RECOVERY_JOURNAL_HEADER_7_0);

 initial_offset = *offset;
 encode_u64_le(buffer, offset, state.journal_start);
 encode_u64_le(buffer, offset, state.logical_blocks_used);
 encode_u64_le(buffer, offset, state.block_map_data_blocks);

 VDO_ASSERT_LOG_ONLY(VDO_RECOVERY_JOURNAL_HEADER_7_0.size == *offset - initial_offset,
       "encoded recovery journal component size must match header size");
}

/**
 * decode_recovery_journal_state_7_0() - Decode the state of a recovery journal saved in a buffer.
 * @buffer: The buffer containing the saved state.
 * @state: A pointer to a recovery journal state to hold the result of a successful decode.
 *
 * Return: VDO_SUCCESS or an error code.
 */

static int __must_check decode_recovery_journal_state_7_0(u8 *buffer, size_t *offset,
         struct recovery_journal_state_7_0 *state)
{
 struct header header;
 int result;
 size_t initial_offset;
 sequence_number_t journal_start;
 block_count_t logical_blocks_used, block_map_data_blocks;

 vdo_decode_header(buffer, offset, &header);
 result = vdo_validate_header(&VDO_RECOVERY_JOURNAL_HEADER_7_0, &header, true,
         __func__);
 if (result != VDO_SUCCESS)
  return result;

 initial_offset = *offset;
 decode_u64_le(buffer, offset, &journal_start);
 decode_u64_le(buffer, offset, &logical_blocks_used);
 decode_u64_le(buffer, offset, &block_map_data_blocks);

 result = VDO_ASSERT(VDO_RECOVERY_JOURNAL_HEADER_7_0.size == *offset - initial_offset,
       "decoded recovery journal component size must match header size");
 if (result != VDO_SUCCESS)
  return result;

 *state = (struct recovery_journal_state_7_0) {
  .journal_start = journal_start,
  .logical_blocks_used = logical_blocks_used,
  .block_map_data_blocks = block_map_data_blocks,
 };

 return VDO_SUCCESS;
}

/**
 * vdo_get_journal_operation_name() - Get the name of a journal operation.
 * @operation: The operation to name.
 *
 * Return: The name of the operation.
 */

const char *vdo_get_journal_operation_name(enum journal_operation operation)
{
 switch (operation) {
 case VDO_JOURNAL_DATA_REMAPPING:
  return "data remapping";

 case VDO_JOURNAL_BLOCK_MAP_REMAPPING:
  return "block map remapping";

 default:
  return "unknown journal operation";
 }
}

/**
 * encode_slab_depot_state_2_0() - Encode the state of a slab depot into a buffer.
 */

static void encode_slab_depot_state_2_0(u8 *buffer, size_t *offset,
     struct slab_depot_state_2_0 state)
{
 size_t initial_offset;

 vdo_encode_header(buffer, offset, &VDO_SLAB_DEPOT_HEADER_2_0);

 initial_offset = *offset;
 encode_u64_le(buffer, offset, state.slab_config.slab_blocks);
 encode_u64_le(buffer, offset, state.slab_config.data_blocks);
 encode_u64_le(buffer, offset, state.slab_config.reference_count_blocks);
 encode_u64_le(buffer, offset, state.slab_config.slab_journal_blocks);
 encode_u64_le(buffer, offset, state.slab_config.slab_journal_flushing_threshold);
 encode_u64_le(buffer, offset, state.slab_config.slab_journal_blocking_threshold);
 encode_u64_le(buffer, offset, state.slab_config.slab_journal_scrubbing_threshold);
 encode_u64_le(buffer, offset, state.first_block);
 encode_u64_le(buffer, offset, state.last_block);
 buffer[(*offset)++] = state.zone_count;

 VDO_ASSERT_LOG_ONLY(VDO_SLAB_DEPOT_HEADER_2_0.size == *offset - initial_offset,
       "encoded block map component size must match header size");
}

/**
 * decode_slab_depot_state_2_0() - Decode slab depot component state version 2.0 from a buffer.
 *
 * Return: VDO_SUCCESS or an error code.
 */

static int decode_slab_depot_state_2_0(u8 *buffer, size_t *offset,
           struct slab_depot_state_2_0 *state)
{
 struct header header;
 int result;
 size_t initial_offset;
 struct slab_config slab_config;
 block_count_t count;
 physical_block_number_t first_block, last_block;
 zone_count_t zone_count;

 vdo_decode_header(buffer, offset, &header);
 result = vdo_validate_header(&VDO_SLAB_DEPOT_HEADER_2_0, &header, true,
         __func__);
 if (result != VDO_SUCCESS)
  return result;

 initial_offset = *offset;
 decode_u64_le(buffer, offset, &count);
 slab_config.slab_blocks = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.data_blocks = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.reference_count_blocks = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.slab_journal_blocks = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.slab_journal_flushing_threshold = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.slab_journal_blocking_threshold = count;

 decode_u64_le(buffer, offset, &count);
 slab_config.slab_journal_scrubbing_threshold = count;

 decode_u64_le(buffer, offset, &first_block);
 decode_u64_le(buffer, offset, &last_block);
 zone_count = buffer[(*offset)++];

 result = VDO_ASSERT(VDO_SLAB_DEPOT_HEADER_2_0.size == *offset - initial_offset,
       "decoded slab depot component size must match header size");
 if (result != VDO_SUCCESS)
  return result;

 *state = (struct slab_depot_state_2_0) {
  .slab_config = slab_config,
  .first_block = first_block,
  .last_block = last_block,
  .zone_count = zone_count,
 };

 return VDO_SUCCESS;
}

/**
 * vdo_configure_slab_depot() - Configure the slab depot.
 * @partition: The slab depot partition
 * @slab_config: The configuration of a single slab.
 * @zone_count: The number of zones the depot will use.
 * @state: The state structure to be configured.
 *
 * Configures the slab_depot for the specified storage capacity, finding the number of data blocks
 * that will fit and still leave room for the depot metadata, then return the saved state for that
 * configuration.
 *
 * Return: VDO_SUCCESS or an error code.
 */

int vdo_configure_slab_depot(const struct partition *partition,
        struct slab_config slab_config, zone_count_t zone_count,
        struct slab_depot_state_2_0 *state)
{
 block_count_t total_slab_blocks, total_data_blocks;
 size_t slab_count;
 physical_block_number_t last_block;
 block_count_t slab_size = slab_config.slab_blocks;

 vdo_log_debug("slabDepot %s(block_count=%llu, first_block=%llu, slab_size=%llu, zone_count=%u)",
        __func__, (unsigned long long) partition->count,
        (unsigned long long) partition->offset,
        (unsigned long long) slab_size, zone_count);

 /* We do not allow runt slabs, so we waste up to a slab's worth. */
 slab_count = (partition->count / slab_size);
 if (slab_count == 0)
  return VDO_NO_SPACE;

 if (slab_count > MAX_VDO_SLABS)
  return VDO_TOO_MANY_SLABS;

 total_slab_blocks = slab_count * slab_config.slab_blocks;
 total_data_blocks = slab_count * slab_config.data_blocks;
 last_block = partition->offset + total_slab_blocks;

 *state = (struct slab_depot_state_2_0) {
  .slab_config = slab_config,
  .first_block = partition->offset,
  .last_block = last_block,
  .zone_count = zone_count,
 };

 vdo_log_debug("slab_depot last_block=%llu, total_data_blocks=%llu, slab_count=%zu, left_over=%llu",
        (unsigned long long) last_block,
        (unsigned long long) total_data_blocks, slab_count,
        (unsigned long long) (partition->count - (last_block - partition->offset)));

 return VDO_SUCCESS;
}

/**
 * vdo_configure_slab() - Measure and initialize the configuration to use for each slab.
 * @slab_size: The number of blocks per slab.
 * @slab_journal_blocks: The number of blocks for the slab journal.
 * @slab_config: The slab configuration to initialize.
 *
 * Return: VDO_SUCCESS or an error code.
 */

int vdo_configure_slab(block_count_t slab_size, block_count_t slab_journal_blocks,
         struct slab_config *slab_config)
{
 block_count_t ref_blocks, meta_blocks, data_blocks;
 block_count_t flushing_threshold, remaining, blocking_threshold;
 block_count_t minimal_extra_space, scrubbing_threshold;

 if (slab_journal_blocks >= slab_size)
  return VDO_BAD_CONFIGURATION;

 /*
 * This calculation should technically be a recurrence, but the total number of metadata
 * blocks is currently less than a single block of ref_counts, so we'd gain at most one
 * data block in each slab with more iteration.
 */

 ref_blocks = vdo_get_saved_reference_count_size(slab_size - slab_journal_blocks);
 meta_blocks = (ref_blocks + slab_journal_blocks);

 /* Make sure configured slabs are not too small. */
 if (meta_blocks >= slab_size)
  return VDO_BAD_CONFIGURATION;

 data_blocks = slab_size - meta_blocks;

 /*
 * Configure the slab journal thresholds. The flush threshold is 168 of 224 blocks in
 * production, or 3/4ths, so we use this ratio for all sizes.
 */

 flushing_threshold = ((slab_journal_blocks * 3) + 3) / 4;
 /*
 * The blocking threshold should be far enough from the flushing threshold to not produce
 * delays, but far enough from the end of the journal to allow multiple successive recovery
 * failures.
 */

 remaining = slab_journal_blocks - flushing_threshold;
 blocking_threshold = flushing_threshold + ((remaining * 5) / 7);
 /* The scrubbing threshold should be at least 2048 entries before the end of the journal. */
 minimal_extra_space = 1 + (MAXIMUM_VDO_USER_VIOS / VDO_SLAB_JOURNAL_FULL_ENTRIES_PER_BLOCK);
 scrubbing_threshold = blocking_threshold;
 if (slab_journal_blocks > minimal_extra_space)
  scrubbing_threshold = slab_journal_blocks - minimal_extra_space;
 if (blocking_threshold > scrubbing_threshold)
  blocking_threshold = scrubbing_threshold;

 *slab_config = (struct slab_config) {
  .slab_blocks = slab_size,
  .data_blocks = data_blocks,
  .reference_count_blocks = ref_blocks,
  .slab_journal_blocks = slab_journal_blocks,
  .slab_journal_flushing_threshold = flushing_threshold,
  .slab_journal_blocking_threshold = blocking_threshold,
  .slab_journal_scrubbing_threshold = scrubbing_threshold};
 return VDO_SUCCESS;
}

/**
 * vdo_decode_slab_journal_entry() - Decode a slab journal entry.
 * @block: The journal block holding the entry.
 * @entry_count: The number of the entry.
 *
 * Return: The decoded entry.
 */

struct slab_journal_entry vdo_decode_slab_journal_entry(struct packed_slab_journal_block *block,
       journal_entry_count_t entry_count)
{
 struct slab_journal_entry entry =
  vdo_unpack_slab_journal_entry(&block->payload.entries[entry_count]);

 if (block->header.has_block_map_increments &&
     ((block->payload.full_entries.entry_types[entry_count / 8] &
       ((u8) 1 << (entry_count % 8))) != 0))
  entry.operation = VDO_JOURNAL_BLOCK_MAP_REMAPPING;

 return entry;
}

/**
 * allocate_partition() - Allocate a partition and add it to a layout.
 * @layout: The layout containing the partition.
 * @id: The id of the partition.
 * @offset: The offset into the layout at which the partition begins.
 * @size: The size of the partition in blocks.
 *
 * Return: VDO_SUCCESS or an error.
 */

static int allocate_partition(struct layout *layout, u8 id,
         physical_block_number_t offset, block_count_t size)
{
 struct partition *partition;
 int result;

 result = vdo_allocate(1, struct partition, __func__, &partition);
 if (result != VDO_SUCCESS)
  return result;

 partition->id = id;
 partition->offset = offset;
 partition->count = size;
 partition->next = layout->head;
 layout->head = partition;

 return VDO_SUCCESS;
}

/**
 * make_partition() - Create a new partition from the beginning or end of the unused space in a
 *                    layout.
 * @layout: The layout.
 * @id: The id of the partition to make.
 * @size: The number of blocks to carve out; if 0, all remaining space will be used.
 * @beginning: True if the partition should start at the beginning of the unused space.
 *
 * Return: A success or error code, particularly VDO_NO_SPACE if there are fewer than size blocks
 *         remaining.
 */

static int __must_check make_partition(struct layout *layout, enum partition_id id,
           block_count_t size, bool beginning)
{
 int result;
 physical_block_number_t offset;
 block_count_t free_blocks = layout->last_free - layout->first_free;

 if (size == 0) {
  if (free_blocks == 0)
   return VDO_NO_SPACE;
  size = free_blocks;
 } else if (size > free_blocks) {
  return VDO_NO_SPACE;
 }

 result = vdo_get_partition(layout, id, NULL);
 if (result != VDO_UNKNOWN_PARTITION)
  return VDO_PARTITION_EXISTS;

 offset = beginning ? layout->first_free : (layout->last_free - size);

 result = allocate_partition(layout, id, offset, size);
 if (result != VDO_SUCCESS)
  return result;

 layout->num_partitions++;
 if (beginning)
  layout->first_free += size;
 else
  layout->last_free = layout->last_free - size;

 return VDO_SUCCESS;
}

/**
 * vdo_initialize_layout() - Lay out the partitions of a vdo.
 * @size: The entire size of the vdo.
 * @offset: The start of the layout on the underlying storage in blocks.
 * @block_map_blocks: The size of the block map partition.
 * @journal_blocks: The size of the journal partition.
 * @summary_blocks: The size of the slab summary partition.
 * @layout: The layout to initialize.
 *
 * Return: VDO_SUCCESS or an error.
 */

int vdo_initialize_layout(block_count_t size, physical_block_number_t offset,
     block_count_t block_map_blocks, block_count_t journal_blocks,
     block_count_t summary_blocks, struct layout *layout)
{
 int result;
 block_count_t necessary_size =
  (offset + block_map_blocks + journal_blocks + summary_blocks);

 if (necessary_size > size)
  return vdo_log_error_strerror(VDO_NO_SPACE,
           "Not enough space to make a VDO");

 *layout = (struct layout) {
  .start = offset,
  .size = size,
  .first_free = offset,
  .last_free = size,
  .num_partitions = 0,
  .head = NULL,
 };

 result = make_partition(layout, VDO_BLOCK_MAP_PARTITION, block_map_blocks, true);
 if (result != VDO_SUCCESS) {
  vdo_uninitialize_layout(layout);
  return result;
 }

 result = make_partition(layout, VDO_SLAB_SUMMARY_PARTITION, summary_blocks,
    false);
 if (result != VDO_SUCCESS) {
  vdo_uninitialize_layout(layout);
  return result;
 }

 result = make_partition(layout, VDO_RECOVERY_JOURNAL_PARTITION, journal_blocks,
    false);
 if (result != VDO_SUCCESS) {
  vdo_uninitialize_layout(layout);
  return result;
 }

 result = make_partition(layout, VDO_SLAB_DEPOT_PARTITION, 0, true);
 if (result != VDO_SUCCESS)
  vdo_uninitialize_layout(layout);

 return result;
}

/**
 * vdo_uninitialize_layout() - Clean up a layout.
 * @layout: The layout to clean up.
 *
 * All partitions created by this layout become invalid pointers.
 */

void vdo_uninitialize_layout(struct layout *layout)
{
 while (layout->head != NULL) {
  struct partition *part = layout->head;

  layout->head = part->next;
  vdo_free(part);
 }

 memset(layout, 0, sizeof(struct layout));
}

/**
 * vdo_get_partition() - Get a partition by id.
 * @layout: The layout from which to get a partition.
 * @id: The id of the partition.
 * @partition_ptr: A pointer to hold the partition.
 *
 * Return: VDO_SUCCESS or an error.
 */

int vdo_get_partition(struct layout *layout, enum partition_id id,
        struct partition **partition_ptr)
{
 struct partition *partition;

 for (partition = layout->head; partition != NULL; partition = partition->next) {
  if (partition->id == id) {
   if (partition_ptr != NULL)
    *partition_ptr = partition;
   return VDO_SUCCESS;
  }
 }

 return VDO_UNKNOWN_PARTITION;
}

/**
 * vdo_get_known_partition() - Get a partition by id from a validated layout.
 * @layout: The layout from which to get a partition.
 * @id: The id of the partition.
 *
 * Return: the partition
 */

struct partition *vdo_get_known_partition(struct layout *layout, enum partition_id id)
{
 struct partition *partition;
 int result = vdo_get_partition(layout, id, &partition);

 VDO_ASSERT_LOG_ONLY(result == VDO_SUCCESS, "layout has expected partition: %u", id);

 return partition;
}

static void encode_layout(u8 *buffer, size_t *offset, const struct layout *layout)
{
 const struct partition *partition;
 size_t initial_offset;
 struct header header = VDO_LAYOUT_HEADER_3_0;

 BUILD_BUG_ON(sizeof(enum partition_id) != sizeof(u8));
 VDO_ASSERT_LOG_ONLY(layout->num_partitions <= U8_MAX,
       "layout partition count must fit in a byte");

 vdo_encode_header(buffer, offset, &header);

 initial_offset = *offset;
 encode_u64_le(buffer, offset, layout->first_free);
 encode_u64_le(buffer, offset, layout->last_free);
 buffer[(*offset)++] = layout->num_partitions;

 VDO_ASSERT_LOG_ONLY(sizeof(struct layout_3_0) == *offset - initial_offset,
       "encoded size of a layout header must match structure");

 for (partition = layout->head; partition != NULL; partition = partition->next) {
  buffer[(*offset)++] = partition->id;
  encode_u64_le(buffer, offset, partition->offset);
  /* This field only exists for backwards compatibility */
  encode_u64_le(buffer, offset, 0);
  encode_u64_le(buffer, offset, partition->count);
 }

 VDO_ASSERT_LOG_ONLY(header.size == *offset - initial_offset,
       "encoded size of a layout must match header size");
}

static int decode_layout(u8 *buffer, size_t *offset, physical_block_number_t start,
    block_count_t size, struct layout *layout)
{
 struct header header;
 struct layout_3_0 layout_header;
 struct partition *partition;
 size_t initial_offset;
 physical_block_number_t first_free, last_free;
 u8 partition_count;
 u8 i;
 int result;

 vdo_decode_header(buffer, offset, &header);
 /* Layout is variable size, so only do a minimum size check here. */
 result = vdo_validate_header(&VDO_LAYOUT_HEADER_3_0, &header, false, __func__);
 if (result != VDO_SUCCESS)
  return result;

 initial_offset = *offset;
 decode_u64_le(buffer, offset, &first_free);
 decode_u64_le(buffer, offset, &last_free);
 partition_count = buffer[(*offset)++];
 layout_header = (struct layout_3_0) {
  .first_free = first_free,
  .last_free = last_free,
  .partition_count = partition_count,
 };

 result = VDO_ASSERT(sizeof(struct layout_3_0) == *offset - initial_offset,
       "decoded size of a layout header must match structure");
 if (result != VDO_SUCCESS)
  return result;

 layout->start = start;
 layout->size = size;
 layout->first_free = layout_header.first_free;
 layout->last_free = layout_header.last_free;
 layout->num_partitions = layout_header.partition_count;

 if (layout->num_partitions > VDO_PARTITION_COUNT) {
  return vdo_log_error_strerror(VDO_UNKNOWN_PARTITION,
           "layout has extra partitions");
 }

 for (i = 0; i < layout->num_partitions; i++) {
  u8 id;
  u64 partition_offset, count;

  id = buffer[(*offset)++];
  decode_u64_le(buffer, offset, &partition_offset);
  *offset += sizeof(u64);
  decode_u64_le(buffer, offset, &count);

  result = allocate_partition(layout, id, partition_offset, count);
  if (result != VDO_SUCCESS) {
   vdo_uninitialize_layout(layout);
   return result;
  }
 }

 /* Validate that the layout has all (and only) the required partitions */
 for (i = 0; i < VDO_PARTITION_COUNT; i++) {
  result = vdo_get_partition(layout, REQUIRED_PARTITIONS[i], &partition);
  if (result != VDO_SUCCESS) {
   vdo_uninitialize_layout(layout);
   return vdo_log_error_strerror(result,
            "layout is missing required partition %u",
            REQUIRED_PARTITIONS[i]);
  }

  start += partition->count;
 }

 if (start != size) {
  vdo_uninitialize_layout(layout);
  return vdo_log_error_strerror(UDS_BAD_STATE,
           "partitions do not cover the layout");
 }

 return VDO_SUCCESS;
}

/**
 * pack_vdo_config() - Convert a vdo_config to its packed on-disk representation.
 * @config: The vdo config to convert.
 *
 * Return: The platform-independent representation of the config.
 */

static struct packed_vdo_config pack_vdo_config(struct vdo_config config)
{
 return (struct packed_vdo_config) {
  .logical_blocks = __cpu_to_le64(config.logical_blocks),
  .physical_blocks = __cpu_to_le64(config.physical_blocks),
  .slab_size = __cpu_to_le64(config.slab_size),
  .recovery_journal_size = __cpu_to_le64(config.recovery_journal_size),
  .slab_journal_blocks = __cpu_to_le64(config.slab_journal_blocks),
 };
}

/**
 * pack_vdo_component() - Convert a vdo_component to its packed on-disk representation.
 * @component: The VDO component data to convert.
 *
 * Return: The platform-independent representation of the component.
 */

static struct packed_vdo_component_41_0 pack_vdo_component(const struct vdo_component component)
{
 return (struct packed_vdo_component_41_0) {
  .state = __cpu_to_le32(component.state),
  .complete_recoveries = __cpu_to_le64(component.complete_recoveries),
  .read_only_recoveries = __cpu_to_le64(component.read_only_recoveries),
  .config = pack_vdo_config(component.config),
  .nonce = __cpu_to_le64(component.nonce),
 };
}

static void encode_vdo_component(u8 *buffer, size_t *offset,
     struct vdo_component component)
{
 struct packed_vdo_component_41_0 packed;

 encode_version_number(buffer, offset, VDO_COMPONENT_DATA_41_0);
 packed = pack_vdo_component(component);
 memcpy(buffer + *offset, &packed, sizeof(packed));
 *offset += sizeof(packed);
}

/**
 * unpack_vdo_config() - Convert a packed_vdo_config to its native in-memory representation.
 * @config: The packed vdo config to convert.
 *
 * Return: The native in-memory representation of the vdo config.
 */

static struct vdo_config unpack_vdo_config(struct packed_vdo_config config)
{
 return (struct vdo_config) {
  .logical_blocks = __le64_to_cpu(config.logical_blocks),
  .physical_blocks = __le64_to_cpu(config.physical_blocks),
  .slab_size = __le64_to_cpu(config.slab_size),
  .recovery_journal_size = __le64_to_cpu(config.recovery_journal_size),
  .slab_journal_blocks = __le64_to_cpu(config.slab_journal_blocks),
 };
}

/**
 * unpack_vdo_component_41_0() - Convert a packed_vdo_component_41_0 to its native in-memory
 *  representation.
 * @component: The packed vdo component data to convert.
 *
 * Return: The native in-memory representation of the component.
 */

static struct vdo_component unpack_vdo_component_41_0(struct packed_vdo_component_41_0 component)
{
 return (struct vdo_component) {
  .state = __le32_to_cpu(component.state),
  .complete_recoveries = __le64_to_cpu(component.complete_recoveries),
  .read_only_recoveries = __le64_to_cpu(component.read_only_recoveries),
  .config = unpack_vdo_config(component.config),
  .nonce = __le64_to_cpu(component.nonce),
 };
}

/**
 * decode_vdo_component() - Decode the component data for the vdo itself out of the super block.
 *
 * Return: VDO_SUCCESS or an error.
 */

static int decode_vdo_component(u8 *buffer, size_t *offset, struct vdo_component *component)
{
 struct version_number version;
 struct packed_vdo_component_41_0 packed;
 int result;

 decode_version_number(buffer, offset, &version);
 result = validate_version(version, VDO_COMPONENT_DATA_41_0,
      "VDO component data");
 if (result != VDO_SUCCESS)
  return result;

 memcpy(&packed, buffer + *offset, sizeof(packed));
 *offset += sizeof(packed);
 *component = unpack_vdo_component_41_0(packed);
 return VDO_SUCCESS;
}

/**
 * vdo_validate_config() - Validate constraints on a VDO config.
 * @config: The VDO config.
 * @physical_block_count: The minimum block count of the underlying storage.
 * @logical_block_count: The expected logical size of the VDO, or 0 if the logical size may be
 *  unspecified.
 *
 * Return: A success or error code.
 */

int vdo_validate_config(const struct vdo_config *config,
   block_count_t physical_block_count,
   block_count_t logical_block_count)
{
 struct slab_config slab_config;
 int result;

 result = VDO_ASSERT(config->slab_size > 0, "slab size unspecified");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(is_power_of_2(config->slab_size),
       "slab size must be a power of two");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(config->slab_size <= (1 << MAX_VDO_SLAB_BITS),
       "slab size must be less than or equal to 2^%d",
       MAX_VDO_SLAB_BITS);
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(config->slab_journal_blocks <= config->slab_size,
       "slab journal size is within expected bound");
 if (result != VDO_SUCCESS)
  return result;

 result = vdo_configure_slab(config->slab_size, config->slab_journal_blocks,
        &slab_config);
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT((slab_config.data_blocks >= 1),
       "slab must be able to hold at least one block");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(config->physical_blocks > 0, "physical blocks unspecified");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(config->physical_blocks <= MAXIMUM_VDO_PHYSICAL_BLOCKS,
       "physical block count %llu exceeds maximum %llu",
       (unsigned long long) config->physical_blocks,
       (unsigned long long) MAXIMUM_VDO_PHYSICAL_BLOCKS);
 if (result != VDO_SUCCESS)
  return VDO_OUT_OF_RANGE;

 if (physical_block_count != config->physical_blocks) {
  vdo_log_error("A physical size of %llu blocks was specified, not the %llu blocks configured in the vdo super block",
         (unsigned long long) physical_block_count,
         (unsigned long long) config->physical_blocks);
  return VDO_PARAMETER_MISMATCH;
 }

 if (logical_block_count > 0) {
  result = VDO_ASSERT((config->logical_blocks > 0),
        "logical blocks unspecified");
  if (result != VDO_SUCCESS)
   return result;

  if (logical_block_count != config->logical_blocks) {
   vdo_log_error("A logical size of %llu blocks was specified, but that differs from the %llu blocks configured in the vdo super block",
          (unsigned long long) logical_block_count,
          (unsigned long long) config->logical_blocks);
   return VDO_PARAMETER_MISMATCH;
  }
 }

 result = VDO_ASSERT(config->logical_blocks <= MAXIMUM_VDO_LOGICAL_BLOCKS,
       "logical blocks too large");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(config->recovery_journal_size > 0,
       "recovery journal size unspecified");
 if (result != VDO_SUCCESS)
  return result;

 result = VDO_ASSERT(is_power_of_2(config->recovery_journal_size),
       "recovery journal size must be a power of two");
 if (result != VDO_SUCCESS)
  return result;

 return result;
}

/**
 * vdo_destroy_component_states() - Clean up any allocations in a vdo_component_states.
 * @states: The component states to destroy.
 */

void vdo_destroy_component_states(struct vdo_component_states *states)
{
 if (states == NULL)
  return;

 vdo_uninitialize_layout(&states->layout);
}

/**
 * decode_components() - Decode the components now that we know the component data is a version we
 *                       understand.
 * @buffer: The buffer being decoded.
 * @offset: The offset to start decoding from.
 * @geometry: The vdo geometry
 * @states: An object to hold the successfully decoded state.
 *
 * Return: VDO_SUCCESS or an error.
 */

static int __must_check decode_components(u8 *buffer, size_t *offset,
       struct volume_geometry *geometry,
       struct vdo_component_states *states)
{
 int result;

 decode_vdo_component(buffer, offset, &states->vdo);

 result = decode_layout(buffer, offset, vdo_get_data_region_start(*geometry) + 1,
          states->vdo.config.physical_blocks, &states->layout);
 if (result != VDO_SUCCESS)
  return result;

 result = decode_recovery_journal_state_7_0(buffer, offset,
         &states->recovery_journal);
 if (result != VDO_SUCCESS)
  return result;

 result = decode_slab_depot_state_2_0(buffer, offset, &states->slab_depot);
 if (result != VDO_SUCCESS)
  return result;

 result = decode_block_map_state_2_0(buffer, offset, &states->block_map);
 if (result != VDO_SUCCESS)
  return result;

 VDO_ASSERT_LOG_ONLY(*offset == VDO_COMPONENT_DATA_OFFSET + VDO_COMPONENT_DATA_SIZE,
       "All decoded component data was used");
 return VDO_SUCCESS;
}

/**
 * vdo_decode_component_states() - Decode the payload of a super block.
 * @buffer: The buffer containing the encoded super block contents.
 * @geometry: The vdo geometry
 * @states: A pointer to hold the decoded states.
 *
 * Return: VDO_SUCCESS or an error.
 */

int vdo_decode_component_states(u8 *buffer, struct volume_geometry *geometry,
    struct vdo_component_states *states)
{
 int result;
 size_t offset = VDO_COMPONENT_DATA_OFFSET;

 /* This is for backwards compatibility. */
 decode_u32_le(buffer, &offset, &states->unused);

 /* Check the VDO volume version */
 decode_version_number(buffer, &offset, &states->volume_version);
 result = validate_version(VDO_VOLUME_VERSION_67_0, states->volume_version,
      "volume");
 if (result != VDO_SUCCESS)
  return result;

 result = decode_components(buffer, &offset, geometry, states);
 if (result != VDO_SUCCESS)
  vdo_uninitialize_layout(&states->layout);

 return result;
}

/**
 * vdo_validate_component_states() - Validate the decoded super block configuration.
 * @states: The state decoded from the super block.
 * @geometry_nonce: The nonce from the geometry block.
 * @physical_size: The minimum block count of the underlying storage.
 * @logical_size: The expected logical size of the VDO, or 0 if the logical size may be
 *                unspecified.
 *
 * Return: VDO_SUCCESS or an error if the configuration is invalid.
 */

int vdo_validate_component_states(struct vdo_component_states *states,
      nonce_t geometry_nonce, block_count_t physical_size,
      block_count_t logical_size)
{
 if (geometry_nonce != states->vdo.nonce) {
  return vdo_log_error_strerror(VDO_BAD_NONCE,
           "Geometry nonce %llu does not match superblock nonce %llu",
           (unsigned long long) geometry_nonce,
           (unsigned long long) states->vdo.nonce);
 }

 return vdo_validate_config(&states->vdo.config, physical_size, logical_size);
}

/**
 * vdo_encode_component_states() - Encode the state of all vdo components in the super block.
 */

static void vdo_encode_component_states(u8 *buffer, size_t *offset,
     const struct vdo_component_states *states)
{
 /* This is for backwards compatibility. */
 encode_u32_le(buffer, offset, states->unused);
 encode_version_number(buffer, offset, states->volume_version);
 encode_vdo_component(buffer, offset, states->vdo);
 encode_layout(buffer, offset, &states->layout);
 encode_recovery_journal_state_7_0(buffer, offset, states->recovery_journal);
 encode_slab_depot_state_2_0(buffer, offset, states->slab_depot);
 encode_block_map_state_2_0(buffer, offset, states->block_map);

 VDO_ASSERT_LOG_ONLY(*offset == VDO_COMPONENT_DATA_OFFSET + VDO_COMPONENT_DATA_SIZE,
       "All super block component data was encoded");
}

/**
 * vdo_encode_super_block() - Encode a super block into its on-disk representation.
 */

void vdo_encode_super_block(u8 *buffer, struct vdo_component_states *states)
{
 u32 checksum;
 struct header header = SUPER_BLOCK_HEADER_12_0;
 size_t offset = 0;

 header.size += VDO_COMPONENT_DATA_SIZE;
 vdo_encode_header(buffer, &offset, &header);
 vdo_encode_component_states(buffer, &offset, states);

 checksum = vdo_crc32(buffer, offset);
 encode_u32_le(buffer, &offset, checksum);

 /*
 * Even though the buffer is a full block, to avoid the potential corruption from a torn
 * write, the entire encoding must fit in the first sector.
 */

 VDO_ASSERT_LOG_ONLY(offset <= VDO_SECTOR_SIZE,
       "entire superblock must fit in one sector");
}

/**
 * vdo_decode_super_block() - Decode a super block from its on-disk representation.
 */

int vdo_decode_super_block(u8 *buffer)
{
 struct header header;
 int result;
 u32 checksum, saved_checksum;
 size_t offset = 0;

 /* Decode and validate the header. */
 vdo_decode_header(buffer, &offset, &header);
 result = vdo_validate_header(&SUPER_BLOCK_HEADER_12_0, &header, false, __func__);
 if (result != VDO_SUCCESS)
  return result;

 if (header.size > VDO_COMPONENT_DATA_SIZE + sizeof(u32)) {
  /*
 * We can't check release version or checksum until we know the content size, so we
 * have to assume a version mismatch on unexpected values.
 */

  return vdo_log_error_strerror(VDO_UNSUPPORTED_VERSION,
           "super block contents too large: %zu",
           header.size);
 }

 /* Skip past the component data for now, to verify the checksum. */
 offset += VDO_COMPONENT_DATA_SIZE;

 checksum = vdo_crc32(buffer, offset);
 decode_u32_le(buffer, &offset, &saved_checksum);

 result = VDO_ASSERT(offset == VDO_SUPER_BLOCK_FIXED_SIZE + VDO_COMPONENT_DATA_SIZE,
       "must have decoded entire superblock payload");
 if (result != VDO_SUCCESS)
  return result;

 return ((checksum != saved_checksum) ? VDO_CHECKSUM_MISMATCH : VDO_SUCCESS);
}

Messung V0.5
C=94 H=92 G=92

¤ Dauer der Verarbeitung: 0.14 Sekunden  ¤

*© Formatika GbR, Deutschland






Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung und die Messung sind noch experimentell.