zstd/tests/seqgen.c

261 lines
7.8 KiB
C
Raw Normal View History

/*
* Copyright (c) 2017-2020, 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 "seqgen.h"
#include "mem.h"
#include <string.h>
#define MIN(a, b) ((a) < (b) ? (a) : (b))
static const size_t kMatchBytes = 128;
#define SEQ_rotl32(x,r) ((x << r) | (x >> (32 - r)))
fix confusion between unsigned <-> U32 as suggested in #1441. generally U32 and unsigned are the same thing, except when they are not ... case : 32-bit compilation for MIPS (uint32_t == unsigned long) A vast majority of transformation consists in transforming U32 into unsigned. In rare cases, it's the other way around (typically for internal code, such as seeds). Among a few issues this patches solves : - some parameters were declared with type `unsigned` in *.h, but with type `U32` in their implementation *.c . - some parameters have type unsigned*, but the caller user a pointer to U32 instead. These fixes are useful. However, the bulk of changes is about %u formating, which requires unsigned type, but generally receives U32 values instead, often just for brevity (U32 is shorter than unsigned). These changes are generally minor, or even annoying. As a consequence, the amount of code changed is larger than I would expect for such a patch. Testing is also a pain : it requires manually modifying `mem.h`, in order to lie about `U32` and force it to be an `unsigned long` typically. On a 64-bit system, this will break the equivalence unsigned == U32. Unfortunately, it will also break a few static_assert(), controlling structure sizes. So it also requires modifying `debug.h` to make `static_assert()` a noop. And then reverting these changes. So it's inconvenient, and as a consequence, this property is currently not checked during CI tests. Therefore, these problems can emerge again in the future. I wonder if it is worth ensuring proper distinction of U32 != unsigned in CI tests. It's another restriction for coding, adding more frustration during merge tests, since most platforms don't need this distinction (hence contributor will not see it), and while this can matter in theory, the number of platforms impacted seems minimal. Thoughts ?
2018-12-21 16:19:44 -08:00
static BYTE SEQ_randByte(unsigned* src)
{
static const U32 prime1 = 2654435761U;
static const U32 prime2 = 2246822519U;
U32 rand32 = *src;
rand32 *= prime1;
rand32 ^= prime2;
rand32 = SEQ_rotl32(rand32, 13);
*src = rand32;
return (BYTE)(rand32 >> 5);
}
SEQ_stream SEQ_initStream(unsigned seed)
{
SEQ_stream stream;
stream.state = 0;
XXH64_reset(&stream.xxh, 0);
stream.seed = seed;
return stream;
}
/* Generates a single guard byte, then match length + 1 of a different byte,
* then another guard byte.
*/
static size_t SEQ_gen_matchLength(SEQ_stream* stream, unsigned value,
SEQ_outBuffer* out)
{
typedef enum {
ml_first_byte = 0,
ml_match_bytes,
ml_last_byte,
} ml_state;
BYTE* const ostart = (BYTE*)out->dst;
BYTE* const oend = ostart + out->size;
BYTE* op = ostart + out->pos;
switch ((ml_state)stream->state) {
case ml_first_byte:
/* Generate a single byte and pick a different byte for the match */
if (op >= oend) {
stream->bytesLeft = 1;
break;
}
*op = SEQ_randByte(&stream->seed) & 0xFF;
do {
stream->saved = SEQ_randByte(&stream->seed) & 0xFF;
} while (*op == stream->saved);
++op;
/* State transition */
stream->state = ml_match_bytes;
stream->bytesLeft = value + 1;
/* fall-through */
case ml_match_bytes: {
/* Copy matchLength + 1 bytes to the output buffer */
size_t const setLength = MIN(stream->bytesLeft, (size_t)(oend - op));
if (setLength > 0) {
memset(op, stream->saved, setLength);
op += setLength;
stream->bytesLeft -= setLength;
}
if (stream->bytesLeft > 0)
break;
/* State transition */
stream->state = ml_last_byte;
}
/* fall-through */
case ml_last_byte:
/* Generate a single byte and pick a different byte for the match */
if (op >= oend) {
stream->bytesLeft = 1;
break;
}
do {
*op = SEQ_randByte(&stream->seed) & 0xFF;
} while (*op == stream->saved);
++op;
/* State transition */
/* fall-through */
default:
stream->state = 0;
stream->bytesLeft = 0;
break;
}
XXH64_update(&stream->xxh, ostart + out->pos, (op - ostart) - out->pos);
out->pos = op - ostart;
return stream->bytesLeft;
}
/* Saves the current seed then generates kMatchBytes random bytes >= 128.
* Generates literal length - kMatchBytes random bytes < 128.
* Generates another kMatchBytes using the saved seed to generate a match.
* This way the match is easy to find for the compressors.
*/
static size_t SEQ_gen_litLength(SEQ_stream* stream, unsigned value, SEQ_outBuffer* out)
{
typedef enum {
ll_start = 0,
ll_run_bytes,
ll_literals,
ll_run_match,
} ll_state;
BYTE* const ostart = (BYTE*)out->dst;
BYTE* const oend = ostart + out->size;
BYTE* op = ostart + out->pos;
switch ((ll_state)stream->state) {
case ll_start:
stream->state = ll_run_bytes;
stream->saved = stream->seed;
stream->bytesLeft = MIN(kMatchBytes, value);
/* fall-through */
case ll_run_bytes:
while (stream->bytesLeft > 0 && op < oend) {
*op++ = SEQ_randByte(&stream->seed) | 0x80;
--stream->bytesLeft;
}
if (stream->bytesLeft > 0)
break;
/* State transition */
stream->state = ll_literals;
stream->bytesLeft = value - MIN(kMatchBytes, value);
/* fall-through */
case ll_literals:
while (stream->bytesLeft > 0 && op < oend) {
*op++ = SEQ_randByte(&stream->seed) & 0x7F;
--stream->bytesLeft;
}
if (stream->bytesLeft > 0)
break;
/* State transition */
stream->state = ll_run_match;
stream->bytesLeft = MIN(kMatchBytes, value);
/* fall-through */
case ll_run_match: {
while (stream->bytesLeft > 0 && op < oend) {
*op++ = SEQ_randByte(&stream->saved) | 0x80;
--stream->bytesLeft;
}
if (stream->bytesLeft > 0)
break;
}
/* fall-through */
default:
stream->state = 0;
stream->bytesLeft = 0;
break;
}
XXH64_update(&stream->xxh, ostart + out->pos, (op - ostart) - out->pos);
out->pos = op - ostart;
return stream->bytesLeft;
}
/* Saves the current seed then generates kMatchBytes random bytes >= 128.
* Generates offset - kMatchBytes of zeros to get a large offset without
* polluting the hash tables.
* Generates another kMatchBytes using the saved seed to generate a with the
* required offset.
*/
static size_t SEQ_gen_offset(SEQ_stream* stream, unsigned value, SEQ_outBuffer* out)
{
typedef enum {
of_start = 0,
of_run_bytes,
of_offset,
of_run_match,
} of_state;
BYTE* const ostart = (BYTE*)out->dst;
BYTE* const oend = ostart + out->size;
BYTE* op = ostart + out->pos;
switch ((of_state)stream->state) {
case of_start:
stream->state = of_run_bytes;
stream->saved = stream->seed;
stream->bytesLeft = MIN(value, kMatchBytes);
/* fall-through */
case of_run_bytes: {
while (stream->bytesLeft > 0 && op < oend) {
*op++ = SEQ_randByte(&stream->seed) | 0x80;
--stream->bytesLeft;
}
if (stream->bytesLeft > 0)
break;
/* State transition */
stream->state = of_offset;
stream->bytesLeft = value - MIN(value, kMatchBytes);
}
/* fall-through */
case of_offset: {
/* Copy matchLength + 1 bytes to the output buffer */
size_t const setLength = MIN(stream->bytesLeft, (size_t)(oend - op));
if (setLength > 0) {
memset(op, 0, setLength);
op += setLength;
stream->bytesLeft -= setLength;
}
if (stream->bytesLeft > 0)
break;
/* State transition */
stream->state = of_run_match;
stream->bytesLeft = MIN(value, kMatchBytes);
}
/* fall-through */
case of_run_match: {
while (stream->bytesLeft > 0 && op < oend) {
*op++ = SEQ_randByte(&stream->saved) | 0x80;
--stream->bytesLeft;
}
if (stream->bytesLeft > 0)
break;
}
/* fall-through */
default:
stream->state = 0;
stream->bytesLeft = 0;
break;
}
XXH64_update(&stream->xxh, ostart + out->pos, (op - ostart) - out->pos);
out->pos = op - ostart;
return stream->bytesLeft;
}
/* Returns the number of bytes left to generate.
* Must pass the same type/value until it returns 0.
*/
size_t SEQ_gen(SEQ_stream* stream, SEQ_gen_type type, unsigned value, SEQ_outBuffer* out)
{
switch (type) {
case SEQ_gen_ml: return SEQ_gen_matchLength(stream, value, out);
case SEQ_gen_ll: return SEQ_gen_litLength(stream, value, out);
case SEQ_gen_of: return SEQ_gen_offset(stream, value, out);
case SEQ_gen_max: /* fall-through */
default: return 0;
}
}
/* Returns the xxhash of the data produced so far */
XXH64_hash_t SEQ_digest(SEQ_stream const* stream)
{
return XXH64_digest(&stream->xxh);
}