linux/lib/zstd/compress/zstd_ldm.c

/*
 * Copyright (c) Yann Collet, Facebook, Inc.
 * All rights reserved.
 *
 * This source code is licensed under both the BSD-style license (found in the
 * LICENSE file in the root directory of this source tree) and the GPLv2 (found
 * in the COPYING file in the root directory of this source tree).
 * You may select, at your option, one of the above-listed licenses.
 */

#include "zstd_ldm.h"

#include "../common/debug.h"
#include <linux/xxhash.h>
#include "zstd_fast.h"          /* ZSTD_fillHashTable() */
#include "zstd_double_fast.h"   /* ZSTD_fillDoubleHashTable() */
#include "zstd_ldm_geartab.h"

#define LDM_BUCKET_SIZE_LOG …
#define LDM_MIN_MATCH_LENGTH …
#define LDM_HASH_RLOG …

ldmRollingHashState_t;

/* ZSTD_ldm_gear_init():
 *
 * Initializes the rolling hash state such that it will honor the
 * settings in params. */
static void ZSTD_ldm_gear_init(ldmRollingHashState_t* state, ldmParams_t const* params)
{ … }

/* ZSTD_ldm_gear_reset()
 * Feeds [data, data + minMatchLength) into the hash without registering any
 * splits. This effectively resets the hash state. This is used when skipping
 * over data, either at the beginning of a block, or skipping sections.
 */
static void ZSTD_ldm_gear_reset(ldmRollingHashState_t* state,
                                BYTE const* data, size_t minMatchLength)
{ … }

/* ZSTD_ldm_gear_feed():
 *
 * Registers in the splits array all the split points found in the first
 * size bytes following the data pointer. This function terminates when
 * either all the data has been processed or LDM_BATCH_SIZE splits are
 * present in the splits array.
 *
 * Precondition: The splits array must not be full.
 * Returns: The number of bytes processed. */
static size_t ZSTD_ldm_gear_feed(ldmRollingHashState_t* state,
                                 BYTE const* data, size_t size,
                                 size_t* splits, unsigned* numSplits)
{ … }

void ZSTD_ldm_adjustParameters(ldmParams_t* params,
                               ZSTD_compressionParameters const* cParams)
{ … }

size_t ZSTD_ldm_getTableSize(ldmParams_t params)
{ … }

size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize)
{ … }

/* ZSTD_ldm_getBucket() :
 *  Returns a pointer to the start of the bucket associated with hash. */
static ldmEntry_t* ZSTD_ldm_getBucket(
        ldmState_t* ldmState, size_t hash, ldmParams_t const ldmParams)
{ … }

/* ZSTD_ldm_insertEntry() :
 *  Insert the entry with corresponding hash into the hash table */
static void ZSTD_ldm_insertEntry(ldmState_t* ldmState,
                                 size_t const hash, const ldmEntry_t entry,
                                 ldmParams_t const ldmParams)
{ … }

/* ZSTD_ldm_countBackwardsMatch() :
 *  Returns the number of bytes that match backwards before pIn and pMatch.
 *
 *  We count only bytes where pMatch >= pBase and pIn >= pAnchor. */
static size_t ZSTD_ldm_countBackwardsMatch(
            const BYTE* pIn, const BYTE* pAnchor,
            const BYTE* pMatch, const BYTE* pMatchBase)
{ … }

/* ZSTD_ldm_countBackwardsMatch_2segments() :
 *  Returns the number of bytes that match backwards from pMatch,
 *  even with the backwards match spanning 2 different segments.
 *
 *  On reaching `pMatchBase`, start counting from mEnd */
static size_t ZSTD_ldm_countBackwardsMatch_2segments(
                    const BYTE* pIn, const BYTE* pAnchor,
                    const BYTE* pMatch, const BYTE* pMatchBase,
                    const BYTE* pExtDictStart, const BYTE* pExtDictEnd)
{ … }

/* ZSTD_ldm_fillFastTables() :
 *
 *  Fills the relevant tables for the ZSTD_fast and ZSTD_dfast strategies.
 *  This is similar to ZSTD_loadDictionaryContent.
 *
 *  The tables for the other strategies are filled within their
 *  block compressors. */
static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,
                                      void const* end)
{ … }

void ZSTD_ldm_fillHashTable(
            ldmState_t* ldmState, const BYTE* ip,
            const BYTE* iend, ldmParams_t const* params)
{ … }


/* ZSTD_ldm_limitTableUpdate() :
 *
 *  Sets cctx->nextToUpdate to a position corresponding closer to anchor
 *  if it is far way
 *  (after a long match, only update tables a limited amount). */
static void ZSTD_ldm_limitTableUpdate(ZSTD_matchState_t* ms, const BYTE* anchor)
{ … }

static size_t ZSTD_ldm_generateSequences_internal(
        ldmState_t* ldmState, rawSeqStore_t* rawSeqStore,
        ldmParams_t const* params, void const* src, size_t srcSize)
{ … }

/*! ZSTD_ldm_reduceTable() :
 *  reduce table indexes by `reducerValue` */
static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size,
                                 U32 const reducerValue)
{ … }

size_t ZSTD_ldm_generateSequences(
        ldmState_t* ldmState, rawSeqStore_t* sequences,
        ldmParams_t const* params, void const* src, size_t srcSize)
{ … }

void
ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize, U32 const minMatch)
{ … }

/*
 * If the sequence length is longer than remaining then the sequence is split
 * between this block and the next.
 *
 * Returns the current sequence to handle, or if the rest of the block should
 * be literals, it returns a sequence with offset == 0.
 */
static rawSeq maybeSplitSequence(rawSeqStore_t* rawSeqStore,
                                 U32 const remaining, U32 const minMatch)
{ … }

void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) { … }

size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
    ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
    ZSTD_paramSwitch_e useRowMatchFinder,
    void const* src, size_t srcSize)
{ … }