2016-07-26 12:41:59 +00:00
|
|
|
/* NOLINT(build/header_guard) */
|
|
|
|
/* Copyright 2016 Google Inc. All Rights Reserved.
|
|
|
|
|
|
|
|
Distributed under MIT license.
|
|
|
|
See file LICENSE for detail or copy at https://opensource.org/licenses/MIT
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* template parameters: FN, BUCKET_BITS, NUM_BANKS, BANK_BITS,
|
|
|
|
NUM_LAST_DISTANCES_TO_CHECK */
|
|
|
|
|
|
|
|
/* A (forgetful) hash table to the data seen by the compressor, to
|
|
|
|
help create backward references to previous data.
|
|
|
|
|
|
|
|
Hashes are stored in chains which are bucketed to groups. Group of chains
|
|
|
|
share a storage "bank". When more than "bank size" chain nodes are added,
|
|
|
|
oldest nodes are replaced; this way several chains may share a tail. */
|
|
|
|
|
|
|
|
#define HashForgetfulChain HASHER()
|
|
|
|
|
|
|
|
#define BANK_SIZE (1 << BANK_BITS)
|
|
|
|
|
|
|
|
/* Number of hash buckets. */
|
|
|
|
#define BUCKET_SIZE (1 << BUCKET_BITS)
|
|
|
|
|
|
|
|
#define CAPPED_CHAINS 0
|
|
|
|
|
|
|
|
static BROTLI_INLINE size_t FN(HashTypeLength)(void) { return 4; }
|
|
|
|
static BROTLI_INLINE size_t FN(StoreLookahead)(void) { return 4; }
|
|
|
|
|
|
|
|
/* HashBytes is the function that chooses the bucket to place the address in.*/
|
|
|
|
static BROTLI_INLINE size_t FN(HashBytes)(const uint8_t *data) {
|
2017-12-12 13:33:12 +00:00
|
|
|
const uint32_t h = BROTLI_UNALIGNED_LOAD32LE(data) * kHashMul32;
|
2016-07-26 12:41:59 +00:00
|
|
|
/* The higher bits contain more mixture from the multiplication,
|
|
|
|
so we take our results from there. */
|
|
|
|
return h >> (32 - BUCKET_BITS);
|
|
|
|
}
|
|
|
|
|
|
|
|
typedef struct FN(Slot) {
|
|
|
|
uint16_t delta;
|
|
|
|
uint16_t next;
|
|
|
|
} FN(Slot);
|
|
|
|
|
|
|
|
typedef struct FN(Bank) {
|
|
|
|
FN(Slot) slots[BANK_SIZE];
|
|
|
|
} FN(Bank);
|
|
|
|
|
|
|
|
typedef struct HashForgetfulChain {
|
|
|
|
uint32_t addr[BUCKET_SIZE];
|
|
|
|
uint16_t head[BUCKET_SIZE];
|
|
|
|
/* Truncated hash used for quick rejection of "distance cache" candidates. */
|
|
|
|
uint8_t tiny_hash[65536];
|
|
|
|
FN(Bank) banks[NUM_BANKS];
|
|
|
|
uint16_t free_slot_idx[NUM_BANKS];
|
|
|
|
size_t max_hops;
|
|
|
|
} HashForgetfulChain;
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE HashForgetfulChain* FN(Self)(HasherHandle handle) {
|
|
|
|
return (HashForgetfulChain*)&(GetHasherCommon(handle)[1]);
|
2017-02-06 13:20:43 +00:00
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static void FN(Initialize)(
|
|
|
|
HasherHandle handle, const BrotliEncoderParams* params) {
|
|
|
|
FN(Self)(handle)->max_hops =
|
|
|
|
(params->quality > 6 ? 7u : 8u) << (params->quality - 4);
|
2017-02-06 13:20:43 +00:00
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static void FN(Prepare)(HasherHandle handle, BROTLI_BOOL one_shot,
|
|
|
|
size_t input_size, const uint8_t* data) {
|
|
|
|
HashForgetfulChain* self = FN(Self)(handle);
|
|
|
|
/* Partial preparation is 100 times slower (per socket). */
|
|
|
|
size_t partial_prepare_threshold = BUCKET_SIZE >> 6;
|
|
|
|
if (one_shot && input_size <= partial_prepare_threshold) {
|
|
|
|
size_t i;
|
|
|
|
for (i = 0; i < input_size; ++i) {
|
|
|
|
size_t bucket = FN(HashBytes)(&data[i]);
|
|
|
|
/* See InitEmpty comment. */
|
|
|
|
self->addr[bucket] = 0xCCCCCCCC;
|
|
|
|
self->head[bucket] = 0xCCCC;
|
|
|
|
}
|
|
|
|
} else {
|
2016-07-26 12:41:59 +00:00
|
|
|
/* Fill |addr| array with 0xCCCCCCCC value. Because of wrapping, position
|
|
|
|
processed by hasher never reaches 3GB + 64M; this makes all new chains
|
|
|
|
to be terminated after the first node. */
|
|
|
|
memset(self->addr, 0xCC, sizeof(self->addr));
|
|
|
|
memset(self->head, 0, sizeof(self->head));
|
|
|
|
}
|
|
|
|
memset(self->tiny_hash, 0, sizeof(self->tiny_hash));
|
|
|
|
memset(self->free_slot_idx, 0, sizeof(self->free_slot_idx));
|
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE size_t FN(HashMemAllocInBytes)(
|
|
|
|
const BrotliEncoderParams* params, BROTLI_BOOL one_shot,
|
|
|
|
size_t input_size) {
|
|
|
|
BROTLI_UNUSED(params);
|
|
|
|
BROTLI_UNUSED(one_shot);
|
|
|
|
BROTLI_UNUSED(input_size);
|
|
|
|
return sizeof(HashForgetfulChain);
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Look at 4 bytes at &data[ix & mask]. Compute a hash from these, and prepend
|
|
|
|
node to corresponding chain; also update tiny_hash for current position. */
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE void FN(Store)(HasherHandle BROTLI_RESTRICT handle,
|
2016-07-26 12:41:59 +00:00
|
|
|
const uint8_t* BROTLI_RESTRICT data, const size_t mask, const size_t ix) {
|
2017-03-06 13:22:45 +00:00
|
|
|
HashForgetfulChain* self = FN(Self)(handle);
|
2016-07-26 12:41:59 +00:00
|
|
|
const size_t key = FN(HashBytes)(&data[ix & mask]);
|
|
|
|
const size_t bank = key & (NUM_BANKS - 1);
|
|
|
|
const size_t idx = self->free_slot_idx[bank]++ & (BANK_SIZE - 1);
|
|
|
|
size_t delta = ix - self->addr[key];
|
|
|
|
self->tiny_hash[(uint16_t)ix] = (uint8_t)key;
|
|
|
|
if (delta > 0xFFFF) delta = CAPPED_CHAINS ? 0 : 0xFFFF;
|
|
|
|
self->banks[bank].slots[idx].delta = (uint16_t)delta;
|
|
|
|
self->banks[bank].slots[idx].next = self->head[key];
|
|
|
|
self->addr[key] = (uint32_t)ix;
|
|
|
|
self->head[key] = (uint16_t)idx;
|
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE void FN(StoreRange)(HasherHandle handle,
|
2016-07-26 12:41:59 +00:00
|
|
|
const uint8_t *data, const size_t mask, const size_t ix_start,
|
|
|
|
const size_t ix_end) {
|
|
|
|
size_t i;
|
|
|
|
for (i = ix_start; i < ix_end; ++i) {
|
2017-03-06 13:22:45 +00:00
|
|
|
FN(Store)(handle, data, mask, i);
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE void FN(StitchToPreviousBlock)(HasherHandle handle,
|
2016-07-26 12:41:59 +00:00
|
|
|
size_t num_bytes, size_t position, const uint8_t* ringbuffer,
|
|
|
|
size_t ring_buffer_mask) {
|
|
|
|
if (num_bytes >= FN(HashTypeLength)() - 1 && position >= 3) {
|
|
|
|
/* Prepare the hashes for three last bytes of the last write.
|
|
|
|
These could not be calculated before, since they require knowledge
|
|
|
|
of both the previous and the current block. */
|
2017-03-06 13:22:45 +00:00
|
|
|
FN(Store)(handle, ringbuffer, ring_buffer_mask, position - 3);
|
|
|
|
FN(Store)(handle, ringbuffer, ring_buffer_mask, position - 2);
|
|
|
|
FN(Store)(handle, ringbuffer, ring_buffer_mask, position - 1);
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
static BROTLI_INLINE void FN(PrepareDistanceCache)(
|
|
|
|
HasherHandle handle, int* BROTLI_RESTRICT distance_cache) {
|
|
|
|
BROTLI_UNUSED(handle);
|
|
|
|
PrepareDistanceCache(distance_cache, NUM_LAST_DISTANCES_TO_CHECK);
|
|
|
|
}
|
|
|
|
|
2016-07-26 12:41:59 +00:00
|
|
|
/* Find a longest backward match of &data[cur_ix] up to the length of
|
|
|
|
max_length and stores the position cur_ix in the hash table.
|
|
|
|
|
2017-03-06 13:22:45 +00:00
|
|
|
REQUIRES: FN(PrepareDistanceCache) must be invoked for current distance cache
|
|
|
|
values; if this method is invoked repeatedly with the same distance
|
|
|
|
cache values, it is enough to invoke FN(PrepareDistanceCache) once.
|
|
|
|
|
2016-07-26 12:41:59 +00:00
|
|
|
Does not look for matches longer than max_length.
|
|
|
|
Does not look for matches further away than max_backward.
|
|
|
|
Writes the best match into |out|.
|
2017-07-21 08:07:24 +00:00
|
|
|
|out|->score is updated only if a better match is found. */
|
|
|
|
static BROTLI_INLINE void FN(FindLongestMatch)(HasherHandle handle,
|
2017-03-06 13:22:45 +00:00
|
|
|
const BrotliDictionary* dictionary, const uint16_t* dictionary_hash,
|
|
|
|
const uint8_t* BROTLI_RESTRICT data, const size_t ring_buffer_mask,
|
|
|
|
const int* BROTLI_RESTRICT distance_cache,
|
2016-07-26 12:41:59 +00:00
|
|
|
const size_t cur_ix, const size_t max_length, const size_t max_backward,
|
2017-08-04 08:02:56 +00:00
|
|
|
const size_t gap, HasherSearchResult* BROTLI_RESTRICT out) {
|
2017-03-06 13:22:45 +00:00
|
|
|
HashForgetfulChain* self = FN(Self)(handle);
|
2016-07-26 12:41:59 +00:00
|
|
|
const size_t cur_ix_masked = cur_ix & ring_buffer_mask;
|
|
|
|
/* Don't accept a short copy from far away. */
|
2017-07-21 08:07:24 +00:00
|
|
|
score_t min_score = out->score;
|
2016-07-26 12:41:59 +00:00
|
|
|
score_t best_score = out->score;
|
|
|
|
size_t best_len = out->len;
|
|
|
|
size_t i;
|
|
|
|
const size_t key = FN(HashBytes)(&data[cur_ix_masked]);
|
|
|
|
const uint8_t tiny_hash = (uint8_t)(key);
|
|
|
|
out->len = 0;
|
2017-07-21 08:07:24 +00:00
|
|
|
out->len_code_delta = 0;
|
2016-07-26 12:41:59 +00:00
|
|
|
/* Try last distance first. */
|
|
|
|
for (i = 0; i < NUM_LAST_DISTANCES_TO_CHECK; ++i) {
|
2017-03-06 13:22:45 +00:00
|
|
|
const size_t backward = (size_t)distance_cache[i];
|
2016-07-26 12:41:59 +00:00
|
|
|
size_t prev_ix = (cur_ix - backward);
|
2016-08-22 11:28:22 +00:00
|
|
|
/* For distance code 0 we want to consider 2-byte matches. */
|
2016-07-26 12:41:59 +00:00
|
|
|
if (i > 0 && self->tiny_hash[(uint16_t)prev_ix] != tiny_hash) continue;
|
|
|
|
if (prev_ix >= cur_ix || backward > max_backward) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
prev_ix &= ring_buffer_mask;
|
|
|
|
{
|
|
|
|
const size_t len = FindMatchLengthWithLimit(&data[prev_ix],
|
|
|
|
&data[cur_ix_masked],
|
|
|
|
max_length);
|
|
|
|
if (len >= 2) {
|
2017-03-06 13:22:45 +00:00
|
|
|
score_t score = BackwardReferenceScoreUsingLastDistance(len);
|
2016-07-26 12:41:59 +00:00
|
|
|
if (best_score < score) {
|
2017-03-06 13:22:45 +00:00
|
|
|
if (i != 0) score -= BackwardReferencePenaltyUsingLastDistance(i);
|
|
|
|
if (best_score < score) {
|
|
|
|
best_score = score;
|
|
|
|
best_len = len;
|
|
|
|
out->len = best_len;
|
|
|
|
out->distance = backward;
|
|
|
|
out->score = best_score;
|
|
|
|
}
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
{
|
|
|
|
const size_t bank = key & (NUM_BANKS - 1);
|
|
|
|
size_t backward = 0;
|
|
|
|
size_t hops = self->max_hops;
|
|
|
|
size_t delta = cur_ix - self->addr[key];
|
|
|
|
size_t slot = self->head[key];
|
|
|
|
while (hops--) {
|
|
|
|
size_t prev_ix;
|
|
|
|
size_t last = slot;
|
|
|
|
backward += delta;
|
|
|
|
if (backward > max_backward || (CAPPED_CHAINS && !delta)) break;
|
|
|
|
prev_ix = (cur_ix - backward) & ring_buffer_mask;
|
|
|
|
slot = self->banks[bank].slots[last].next;
|
|
|
|
delta = self->banks[bank].slots[last].delta;
|
|
|
|
if (cur_ix_masked + best_len > ring_buffer_mask ||
|
|
|
|
prev_ix + best_len > ring_buffer_mask ||
|
|
|
|
data[cur_ix_masked + best_len] != data[prev_ix + best_len]) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
{
|
|
|
|
const size_t len = FindMatchLengthWithLimit(&data[prev_ix],
|
|
|
|
&data[cur_ix_masked],
|
|
|
|
max_length);
|
|
|
|
if (len >= 4) {
|
|
|
|
/* Comparing for >= 3 does not change the semantics, but just saves
|
|
|
|
for a few unnecessary binary logarithms in backward reference
|
|
|
|
score, since we are not interested in such short matches. */
|
|
|
|
score_t score = BackwardReferenceScore(len, backward);
|
|
|
|
if (best_score < score) {
|
|
|
|
best_score = score;
|
|
|
|
best_len = len;
|
|
|
|
out->len = best_len;
|
|
|
|
out->distance = backward;
|
|
|
|
out->score = best_score;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2017-03-06 13:22:45 +00:00
|
|
|
FN(Store)(handle, data, ring_buffer_mask, cur_ix);
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
2017-07-21 08:07:24 +00:00
|
|
|
if (out->score == min_score) {
|
|
|
|
SearchInStaticDictionary(dictionary, dictionary_hash,
|
2017-08-04 08:02:56 +00:00
|
|
|
handle, &data[cur_ix_masked], max_length, max_backward + gap, out,
|
2017-03-06 13:22:45 +00:00
|
|
|
BROTLI_FALSE);
|
2016-07-26 12:41:59 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#undef BANK_SIZE
|
|
|
|
#undef BUCKET_SIZE
|
|
|
|
#undef CAPPED_CHAINS
|
|
|
|
|
|
|
|
#undef HashForgetfulChain
|