dilim/CMake
1
0 Fork 0
mirror of https://github.com/Kitware/CMake.git synced 2026-11-06 13:27:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream-zstd' into update-zstd

Browse Source 
* upstream-zstd:
  zstd 2025-02-18 (f8745da6)
This commit is contained in:
Brad King committed 2026-01-22 09:46:14 -05:00
2 parents 968b7e0b2b 9f5d8ba3ac
commit b352f91585
62 files changed
+7366 -3811
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

Utilities/cmzstd/README.md
+30 -16
View File
@@ -5,7 +5,7 @@ targeting real-time compression scenarios at zlib-level and better compression r
It's backed by a very fast entropy stage, provided by [Huff0 and FSE library](https://github.com/Cyan4973/FiniteStateEntropy).
Zstandard's format is stable and documented in [RFC8878](https://datatracker.ietf.org/doc/html/rfc8878). Multiple independent implementations are already available.
This repository represents the reference implementation, provided as an open-source dual [BSD](LICENSE) and [GPLv2](COPYING) licensed **C** library,
This repository represents the reference implementation, provided as an open-source dual [BSD](LICENSE) OR [GPLv2](COPYING) licensed **C** library,
and a command line utility producing and decoding `.zst`, `.gz`, `.xz` and `.lz4` files.
Should your project require another programming language,
a list of known ports and bindings is provided on [Zstandard homepage](https://facebook.github.io/zstd/#other-languages).
@@ -29,10 +29,10 @@ a list of known ports and bindings is provided on [Zstandard homepage](https://f
## Benchmarks
For reference, several fast compression algorithms were tested and compared
on a desktop running Ubuntu 20.04 (`Linux 5.11.0-41-generic`),
with a Core i7-9700K CPU @ 4.9GHz,
on a desktop featuring a Core i7-9700K CPU @ 4.9GHz
and running Ubuntu 20.04 (`Linux ubu20 5.15.0-101-generic`),
using [lzbench], an open-source in-memory benchmark by @inikep
compiled with [gcc] 9.3.0,
compiled with [gcc] 9.4.0,
on the [Silesia compression corpus].
[lzbench]: https://github.com/inikep/lzbench
@@ -41,24 +41,23 @@ on the [Silesia compression corpus].
| Compressor name | Ratio | Compression| Decompress.|
| --------------- | ------| -----------| ---------- |
| **zstd 1.5.1 -1** | 2.887 | 530 MB/s | 1700 MB/s |
| **zstd 1.5.6 -1** | 2.887 | 510 MB/s | 1580 MB/s |
| [zlib] 1.2.11 -1 | 2.743 | 95 MB/s | 400 MB/s |
| brotli 1.0.9 -0 | 2.702 | 395 MB/s | 450 MB/s |
| **zstd 1.5.1 --fast=1** | 2.437 | 600 MB/s | 2150 MB/s |
| **zstd 1.5.1 --fast=3** | 2.239 | 670 MB/s | 2250 MB/s |
| quicklz 1.5.0 -1 | 2.238 | 540 MB/s | 760 MB/s |
| **zstd 1.5.1 --fast=4** | 2.148 | 710 MB/s | 2300 MB/s |
| lzo1x 2.10 -1 | 2.106 | 660 MB/s | 845 MB/s |
| [lz4] 1.9.3 | 2.101 | 740 MB/s | 4500 MB/s |
| lzf 3.6 -1 | 2.077 | 410 MB/s | 830 MB/s |
| snappy 1.1.9 | 2.073 | 550 MB/s | 1750 MB/s |
| brotli 1.0.9 -0 | 2.702 | 395 MB/s | 430 MB/s |
| **zstd 1.5.6 --fast=1** | 2.437 | 545 MB/s | 1890 MB/s |
| **zstd 1.5.6 --fast=3** | 2.239 | 650 MB/s | 2000 MB/s |
| quicklz 1.5.0 -1 | 2.238 | 525 MB/s | 750 MB/s |
| lzo1x 2.10 -1 | 2.106 | 650 MB/s | 825 MB/s |
| [lz4] 1.9.4 | 2.101 | 700 MB/s | 4000 MB/s |
| lzf 3.6 -1 | 2.077 | 420 MB/s | 830 MB/s |
| snappy 1.1.9 | 2.073 | 530 MB/s | 1660 MB/s |
[zlib]: https://www.zlib.net/
[lz4]: https://lz4.github.io/lz4/
The negative compression levels, specified with `--fast=#`,
offer faster compression and decompression speed
at the cost of compression ratio (compared to level 1).
at the cost of compression ratio.
Zstd can also offer stronger compression ratios at the cost of compression speed.
Speed vs Compression trade-off is configurable by small increments.
@@ -185,6 +184,17 @@ You can build and install zstd [vcpkg](https://github.com/Microsoft/vcpkg/) depe
The zstd port in vcpkg is kept up to date by Microsoft team members and community contributors.
If the version is out of date, please [create an issue or pull request](https://github.com/Microsoft/vcpkg) on the vcpkg repository.
### Conan
You can install pre-built binaries for zstd or build it from source using [Conan](https://conan.io/). Use the following command:
```bash
conan install --requires="zstd/[*]" --build=missing
```
The zstd Conan recipe is kept up to date by Conan maintainers and community contributors.
If the version is out of date, please [create an issue or pull request](https://github.com/conan-io/conan-center-index) on the ConanCenterIndex repository.
### Visual Studio (Windows)
Going into `build` directory, you will find additional possibilities:
@@ -198,6 +208,10 @@ Going into `build` directory, you will find additional possibilities:
You can build the zstd binary via buck by executing: `buck build programs:zstd` from the root of the repo.
The output binary will be in `buck-out/gen/programs/`.
### Bazel
You easily can integrate zstd into your Bazel project by using the module hosted on the [Bazel Central Repository](https://registry.bazel.build/modules/zstd).
## Testing
You can run quick local smoke tests by running `make check`.
@@ -213,7 +227,7 @@ Zstandard is considered safe for production environments.
## License
Zstandard is dual-licensed under [BSD](LICENSE) and [GPLv2](COPYING).
Zstandard is dual-licensed under [BSD](LICENSE) OR [GPLv2](COPYING).
## Contributing
Utilities/cmzstd/lib/common/allocations.h
+1 -1
View File
@@ -14,7 +14,7 @@
#define ZSTD_DEPS_NEED_MALLOC
#include "zstd_deps.h" /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
#include "mem.h" /* MEM_STATIC */
#include "compiler.h" /* MEM_STATIC */
#define ZSTD_STATIC_LINKING_ONLY
#include "../zstd.h" /* ZSTD_customMem */
Utilities/cmzstd/lib/common/bits.h
+92 -87
View File
@@ -28,27 +28,29 @@ MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
{
assert(val != 0);
# if defined(_MSC_VER)
# if STATIC_BMI2 == 1
return (unsigned)_tzcnt_u32(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)__builtin_ctz(val);
# else
return ZSTD_countTrailingZeros32_fallback(val);
# endif
#if defined(_MSC_VER)
# if STATIC_BMI2
return (unsigned)_tzcnt_u32(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward(&r, val);
return (unsigned)r;
} else {
__assume(0); /* Should not reach this code path */
}
# endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)__builtin_ctz(val);
#elif defined(__ICCARM__)
return (unsigned)__builtin_ctz(val);
#else
return ZSTD_countTrailingZeros32_fallback(val);
#endif
}
MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val) {
MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val)
{
assert(val != 0);
{
static const U32 DeBruijnClz[32] = {0, 9, 1, 10, 13, 21, 2, 29,
@@ -67,86 +69,89 @@ MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val) {
MEM_STATIC unsigned ZSTD_countLeadingZeros32(U32 val)
{
assert(val != 0);
# if defined(_MSC_VER)
# if STATIC_BMI2 == 1
return (unsigned)_lzcnt_u32(val);
# else
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, val);
return (unsigned)(31 - r);
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)__builtin_clz(val);
# else
return ZSTD_countLeadingZeros32_fallback(val);
# endif
#if defined(_MSC_VER)
# if STATIC_BMI2
return (unsigned)_lzcnt_u32(val);
# else
if (val != 0) {
unsigned long r;
_BitScanReverse(&r, val);
return (unsigned)(31 - r);
} else {
__assume(0); /* Should not reach this code path */
}
# endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)__builtin_clz(val);
#elif defined(__ICCARM__)
return (unsigned)__builtin_clz(val);
#else
return ZSTD_countLeadingZeros32_fallback(val);
#endif
}
MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
{
assert(val != 0);
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2 == 1
return (unsigned)_tzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward64(&r, val);
return (unsigned)r;
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4) && defined(__LP64__)
return (unsigned)__builtin_ctzll(val);
# else
{
U32 mostSignificantWord = (U32)(val >> 32);
U32 leastSignificantWord = (U32)val;
if (leastSignificantWord == 0) {
return 32 + ZSTD_countTrailingZeros32(mostSignificantWord);
} else {
return ZSTD_countTrailingZeros32(leastSignificantWord);
}
#if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return (unsigned)_tzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanForward64(&r, val);
return (unsigned)r;
} else {
__assume(0); /* Should not reach this code path */
}
# endif
#elif defined(__GNUC__) && (__GNUC__ >= 4) && defined(__LP64__)
return (unsigned)__builtin_ctzll(val);
#elif defined(__ICCARM__)
return (unsigned)__builtin_ctzll(val);
#else
{
U32 mostSignificantWord = (U32)(val >> 32);
U32 leastSignificantWord = (U32)val;
if (leastSignificantWord == 0) {
return 32 + ZSTD_countTrailingZeros32(mostSignificantWord);
} else {
return ZSTD_countTrailingZeros32(leastSignificantWord);
}
# endif
}
#endif
}
MEM_STATIC unsigned ZSTD_countLeadingZeros64(U64 val)
{
assert(val != 0);
# if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2 == 1
return (unsigned)_lzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanReverse64(&r, val);
return (unsigned)(63 - r);
} else {
/* Should not reach this code path */
__assume(0);
}
# endif
# elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)(__builtin_clzll(val));
# else
{
U32 mostSignificantWord = (U32)(val >> 32);
U32 leastSignificantWord = (U32)val;
if (mostSignificantWord == 0) {
return 32 + ZSTD_countLeadingZeros32(leastSignificantWord);
} else {
return ZSTD_countLeadingZeros32(mostSignificantWord);
}
#if defined(_MSC_VER) && defined(_WIN64)
# if STATIC_BMI2
return (unsigned)_lzcnt_u64(val);
# else
if (val != 0) {
unsigned long r;
_BitScanReverse64(&r, val);
return (unsigned)(63 - r);
} else {
__assume(0); /* Should not reach this code path */
}
# endif
#elif defined(__GNUC__) && (__GNUC__ >= 4)
return (unsigned)(__builtin_clzll(val));
#elif defined(__ICCARM__)
return (unsigned)(__builtin_clzll(val));
#else
{
U32 mostSignificantWord = (U32)(val >> 32);
U32 leastSignificantWord = (U32)val;
if (mostSignificantWord == 0) {
return 32 + ZSTD_countLeadingZeros32(leastSignificantWord);
} else {
return ZSTD_countLeadingZeros32(mostSignificantWord);
}
# endif
}
#endif
}
MEM_STATIC unsigned ZSTD_NbCommonBytes(size_t val)
Utilities/cmzstd/lib/common/bitstream.h
+69 -52
View File
@@ -16,9 +16,6 @@
#include <assert.h>
#if defined (__cplusplus)
extern "C" {
#endif
/*
* This API consists of small unitary functions, which must be inlined for best performance.
* Since link-time-optimization is not available for all compilers,
@@ -34,7 +31,6 @@ extern "C" {
#include "error_private.h" /* error codes and messages */
#include "bits.h" /* ZSTD_highbit32 */
/*=========================================
* Target specific
=========================================*/
@@ -54,12 +50,13 @@ extern "C" {
/*-******************************************
* bitStream encoding API (write forward)
********************************************/
typedef size_t BitContainerType;
/* bitStream can mix input from multiple sources.
* A critical property of these streams is that they encode and decode in **reverse** direction.
* So the first bit sequence you add will be the last to be read, like a LIFO stack.
*/
typedef struct {
size_t bitContainer;
BitContainerType bitContainer;
unsigned bitPos;
char* startPtr;
char* ptr;
@@ -67,7 +64,7 @@ typedef struct {
} BIT_CStream_t;
MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC, void* dstBuffer, size_t dstCapacity);
MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC, size_t value, unsigned nbBits);
MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC, BitContainerType value, unsigned nbBits);
MEM_STATIC void BIT_flushBits(BIT_CStream_t* bitC);
MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);
@@ -76,7 +73,7 @@ MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);
* `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code.
*
* bits are first added to a local register.
* Local register is size_t, hence 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
* Local register is BitContainerType, 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
* Writing data into memory is an explicit operation, performed by the flushBits function.
* Hence keep track how many bits are potentially stored into local register to avoid register overflow.
* After a flushBits, a maximum of 7 bits might still be stored into local register.
@@ -93,28 +90,28 @@ MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC);
* bitStream decoding API (read backward)
**********************************************/
typedef struct {
size_t bitContainer;
BitContainerType bitContainer;
unsigned bitsConsumed;
const char* ptr;
const char* start;
const char* limitPtr;
} BIT_DStream_t;
typedef enum { BIT_DStream_unfinished = 0,
BIT_DStream_endOfBuffer = 1,
BIT_DStream_completed = 2,
BIT_DStream_overflow = 3 } BIT_DStream_status; /* result of BIT_reloadDStream() */
/* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */
typedef enum { BIT_DStream_unfinished = 0, /* fully refilled */
BIT_DStream_endOfBuffer = 1, /* still some bits left in bitstream */
BIT_DStream_completed = 2, /* bitstream entirely consumed, bit-exact */
BIT_DStream_overflow = 3 /* user requested more bits than present in bitstream */
} BIT_DStream_status; /* result of BIT_reloadDStream() */
MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize);
MEM_STATIC size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
MEM_STATIC BitContainerType BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD);
MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD);
/* Start by invoking BIT_initDStream().
* A chunk of the bitStream is then stored into a local register.
* Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
* Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (BitContainerType).
* You can then retrieve bitFields stored into the local register, **in reverse order**.
* Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
* A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
@@ -126,7 +123,7 @@ MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* bitD);
/*-****************************************
* unsafe API
******************************************/
MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC, size_t value, unsigned nbBits);
MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC, BitContainerType value, unsigned nbBits);
/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */
MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
@@ -164,10 +161,15 @@ MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC,
return 0;
}
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
FORCE_INLINE_TEMPLATE BitContainerType BIT_getLowerBits(BitContainerType bitContainer, U32 const nbBits)
{
#if defined(STATIC_BMI2) && STATIC_BMI2 == 1 && !defined(ZSTD_NO_INTRINSICS)
return _bzhi_u64(bitContainer, nbBits);
#if STATIC_BMI2 && !defined(ZSTD_NO_INTRINSICS)
# if (defined(__x86_64__) || defined(_M_X64)) && !defined(__ILP32__)
return _bzhi_u64(bitContainer, nbBits);
# else
DEBUG_STATIC_ASSERT(sizeof(bitContainer) == sizeof(U32));
return _bzhi_u32(bitContainer, nbBits);
# endif
#else
assert(nbBits < BIT_MASK_SIZE);
return bitContainer & BIT_mask[nbBits];
@@ -178,7 +180,7 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 co
* can add up to 31 bits into `bitC`.
* Note : does not check for register overflow ! */
MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
size_t value, unsigned nbBits)
BitContainerType value, unsigned nbBits)
{
DEBUG_STATIC_ASSERT(BIT_MASK_SIZE == 32);
assert(nbBits < BIT_MASK_SIZE);
@@ -191,7 +193,7 @@ MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
* works only if `value` is _clean_,
* meaning all high bits above nbBits are 0 */
MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
size_t value, unsigned nbBits)
BitContainerType value, unsigned nbBits)
{
assert((value>>nbBits) == 0);
assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
@@ -238,7 +240,7 @@ MEM_STATIC size_t BIT_closeCStream(BIT_CStream_t* bitC)
BIT_addBitsFast(bitC, 1, 1); /* endMark */
BIT_flushBits(bitC);
if (bitC->ptr >= bitC->endPtr) return 0; /* overflow detected */
return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
return (size_t)(bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
}
@@ -269,22 +271,22 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
bitD->bitContainer = *(const BYTE*)(bitD->start);
switch(srcSize)
{
case 7: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
case 7: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[6]) << (sizeof(bitD->bitContainer)*8 - 16);
ZSTD_FALLTHROUGH;
case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
case 6: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
ZSTD_FALLTHROUGH;
case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
case 5: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
ZSTD_FALLTHROUGH;
case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;
case 4: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[3]) << 24;
ZSTD_FALLTHROUGH;
case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;
case 3: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[2]) << 16;
ZSTD_FALLTHROUGH;
case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) << 8;
case 2: bitD->bitContainer += (BitContainerType)(((const BYTE*)(srcBuffer))[1]) << 8;
ZSTD_FALLTHROUGH;
default: break;
@@ -299,12 +301,12 @@ MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, si
return srcSize;
}
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getUpperBits(size_t bitContainer, U32 const start)
FORCE_INLINE_TEMPLATE BitContainerType BIT_getUpperBits(BitContainerType bitContainer, U32 const start)
{
return bitContainer >> start;
}
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits)
FORCE_INLINE_TEMPLATE BitContainerType BIT_getMiddleBits(BitContainerType bitContainer, U32 const start, U32 const nbBits)
{
U32 const regMask = sizeof(bitContainer)*8 - 1;
/* if start > regMask, bitstream is corrupted, and result is undefined */
@@ -314,7 +316,7 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 c
* such cpus old (pre-Haswell, 2013) and their performance is not of that
* importance.
*/
#if defined(__x86_64__) || defined(_M_X86)
#if defined(__x86_64__) || defined(_M_X64)
return (bitContainer >> (start & regMask)) & ((((U64)1) << nbBits) - 1);
#else
return (bitContainer >> (start & regMask)) & BIT_mask[nbBits];
@@ -327,7 +329,7 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 c
* On 32-bits, maxNbBits==24.
* On 64-bits, maxNbBits==56.
* @return : value extracted */
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_lookBits(const BIT_DStream_t* bitD, U32 nbBits)
FORCE_INLINE_TEMPLATE BitContainerType BIT_lookBits(const BIT_DStream_t* bitD, U32 nbBits)
{
/* arbitrate between double-shift and shift+mask */
#if 1
@@ -343,14 +345,14 @@ MEM_STATIC FORCE_INLINE_ATTR size_t BIT_lookBits(const BIT_DStream_t* bitD, U3
/*! BIT_lookBitsFast() :
* unsafe version; only works if nbBits >= 1 */
MEM_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
MEM_STATIC BitContainerType BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
{
U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
assert(nbBits >= 1);
return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask);
}
MEM_STATIC FORCE_INLINE_ATTR void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
FORCE_INLINE_TEMPLATE void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
{
bitD->bitsConsumed += nbBits;
}
@@ -359,23 +361,38 @@ MEM_STATIC FORCE_INLINE_ATTR void BIT_skipBits(BIT_DStream_t* bitD, U32 nbBits)
* Read (consume) next n bits from local register and update.
* Pay attention to not read more than nbBits contained into local register.
* @return : extracted value. */
MEM_STATIC FORCE_INLINE_ATTR size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
FORCE_INLINE_TEMPLATE BitContainerType BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
{
size_t const value = BIT_lookBits(bitD, nbBits);
BitContainerType const value = BIT_lookBits(bitD, nbBits);
BIT_skipBits(bitD, nbBits);
return value;
}
/*! BIT_readBitsFast() :
* unsafe version; only works if nbBits >= 1 */
MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
MEM_STATIC BitContainerType BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
{
size_t const value = BIT_lookBitsFast(bitD, nbBits);
BitContainerType const value = BIT_lookBitsFast(bitD, nbBits);
assert(nbBits >= 1);
BIT_skipBits(bitD, nbBits);
return value;
}
/*! BIT_reloadDStream_internal() :
* Simple variant of BIT_reloadDStream(), with two conditions:
* 1. bitstream is valid : bitsConsumed <= sizeof(bitD->bitContainer)*8
* 2. look window is valid after shifted down : bitD->ptr >= bitD->start
*/
MEM_STATIC BIT_DStream_status BIT_reloadDStream_internal(BIT_DStream_t* bitD)
{
assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
bitD->ptr -= bitD->bitsConsumed >> 3;
assert(bitD->ptr >= bitD->start);
bitD->bitsConsumed &= 7;
bitD->bitContainer = MEM_readLEST(bitD->ptr);
return BIT_DStream_unfinished;
}
/*! BIT_reloadDStreamFast() :
* Similar to BIT_reloadDStream(), but with two differences:
* 1. bitsConsumed <= sizeof(bitD->bitContainer)*8 must hold!
@@ -386,31 +403,35 @@ MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
{
if (UNLIKELY(bitD->ptr < bitD->limitPtr))
return BIT_DStream_overflow;
assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
bitD->ptr -= bitD->bitsConsumed >> 3;
bitD->bitsConsumed &= 7;
bitD->bitContainer = MEM_readLEST(bitD->ptr);
return BIT_DStream_unfinished;
return BIT_reloadDStream_internal(bitD);
}
/*! BIT_reloadDStream() :
* Refill `bitD` from buffer previously set in BIT_initDStream() .
* This function is safe, it guarantees it will not read beyond src buffer.
* This function is safe, it guarantees it will not never beyond src buffer.
* @return : status of `BIT_DStream_t` internal register.
* when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */
MEM_STATIC FORCE_INLINE_ATTR BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
FORCE_INLINE_TEMPLATE BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
{
if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8)) /* overflow detected, like end of stream */
/* note : once in overflow mode, a bitstream remains in this mode until it's reset */
if (UNLIKELY(bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))) {
static const BitContainerType zeroFilled = 0;
bitD->ptr = (const char*)&zeroFilled; /* aliasing is allowed for char */
/* overflow detected, erroneous scenario or end of stream: no update */
return BIT_DStream_overflow;
}
assert(bitD->ptr >= bitD->start);
if (bitD->ptr >= bitD->limitPtr) {
return BIT_reloadDStreamFast(bitD);
return BIT_reloadDStream_internal(bitD);
}
if (bitD->ptr == bitD->start) {
/* reached end of bitStream => no update */
if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer;
return BIT_DStream_completed;
}
/* start < ptr < limitPtr */
/* start < ptr < limitPtr => cautious update */
{ U32 nbBytes = bitD->bitsConsumed >> 3;
BIT_DStream_status result = BIT_DStream_unfinished;
if (bitD->ptr - nbBytes < bitD->start) {
@@ -432,8 +453,4 @@ MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream)
return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8));
}
#if defined (__cplusplus)
}
#endif
#endif /* BITSTREAM_H_MODULE */
Utilities/cmzstd/lib/common/compiler.h
+148 -42
View File
@@ -11,6 +11,8 @@
#ifndef ZSTD_COMPILER_H
#define ZSTD_COMPILER_H
#include <stddef.h>
#include "portability_macros.h"
/*-*******************************************************
@@ -25,7 +27,7 @@
# define INLINE_KEYWORD
#endif
#if defined(__GNUC__) || defined(__ICCARM__)
#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
# define FORCE_INLINE_ATTR __attribute__((always_inline))
#elif defined(_MSC_VER)
# define FORCE_INLINE_ATTR __forceinline
@@ -51,12 +53,19 @@
# define WIN_CDECL
#endif
/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
# define UNUSED_ATTR __attribute__((unused))
#else
# define UNUSED_ATTR
#endif
/**
* FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
* parameters. They must be inlined for the compiler to eliminate the constant
* branches.
*/
#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR UNUSED_ATTR
/**
* HINT_INLINE is used to help the compiler generate better code. It is *not*
* used for "templates", so it can be tweaked based on the compilers
@@ -71,21 +80,37 @@
#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
# define HINT_INLINE static INLINE_KEYWORD
#else
# define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR
# define HINT_INLINE FORCE_INLINE_TEMPLATE
#endif
/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
/* "soft" inline :
* The compiler is free to select if it's a good idea to inline or not.
* The main objective is to silence compiler warnings
* when a defined function in included but not used.
*
* Note : this macro is prefixed `MEM_` because it used to be provided by `mem.h` unit.
* Updating the prefix is probably preferable, but requires a fairly large codemod,
* since this name is used everywhere.
*/
#ifndef MEM_STATIC /* already defined in Linux Kernel mem.h */
#if defined(__GNUC__)
# define UNUSED_ATTR __attribute__((unused))
# define MEM_STATIC static __inline UNUSED_ATTR
#elif defined(__IAR_SYSTEMS_ICC__)
# define MEM_STATIC static inline UNUSED_ATTR
#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
# define MEM_STATIC static inline
#elif defined(_MSC_VER)
# define MEM_STATIC static __inline
#else
# define UNUSED_ATTR
# define MEM_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */
#endif
#endif
/* force no inlining */
#ifdef _MSC_VER
# define FORCE_NOINLINE static __declspec(noinline)
#else
# if defined(__GNUC__) || defined(__ICCARM__)
# if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
# define FORCE_NOINLINE static __attribute__((__noinline__))
# else
# define FORCE_NOINLINE static
@@ -94,7 +119,7 @@
/* target attribute */
#if defined(__GNUC__) || defined(__ICCARM__)
#if defined(__GNUC__) || defined(__IAR_SYSTEMS_ICC__)
# define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
#else
# define TARGET_ATTRIBUTE(target)
@@ -109,10 +134,10 @@
/* prefetch
* can be disabled, by declaring NO_PREFETCH build macro */
#if defined(NO_PREFETCH)
# define PREFETCH_L1(ptr) (void)(ptr) /* disabled */
# define PREFETCH_L2(ptr) (void)(ptr) /* disabled */
# define PREFETCH_L1(ptr) do { (void)(ptr); } while (0) /* disabled */
# define PREFETCH_L2(ptr) do { (void)(ptr); } while (0) /* disabled */
#else
# if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) /* _mm_prefetch() is not defined outside of x86/x64 */
# if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) && !defined(_M_ARM64EC) /* _mm_prefetch() is not defined outside of x86/x64 */
# include <mmintrin.h> /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */
# define PREFETCH_L1(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0)
# define PREFETCH_L2(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T1)
@@ -120,24 +145,25 @@
# define PREFETCH_L1(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)
# define PREFETCH_L2(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)
# elif defined(__aarch64__)
# define PREFETCH_L1(ptr) __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr)))
# define PREFETCH_L2(ptr) __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr)))
# define PREFETCH_L1(ptr) do { __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr))); } while (0)
# define PREFETCH_L2(ptr) do { __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr))); } while (0)
# else
# define PREFETCH_L1(ptr) (void)(ptr) /* disabled */
# define PREFETCH_L2(ptr) (void)(ptr) /* disabled */
# define PREFETCH_L1(ptr) do { (void)(ptr); } while (0) /* disabled */
# define PREFETCH_L2(ptr) do { (void)(ptr); } while (0) /* disabled */
# endif
#endif /* NO_PREFETCH */
#define CACHELINE_SIZE 64
#define PREFETCH_AREA(p, s) { \
const char* const _ptr = (const char*)(p); \
size_t const _size = (size_t)(s); \
size_t _pos; \
for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \
PREFETCH_L2(_ptr + _pos); \
} \
}
#define PREFETCH_AREA(p, s) \
do { \
const char* const _ptr = (const char*)(p); \
size_t const _size = (size_t)(s); \
size_t _pos; \
for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \
PREFETCH_L2(_ptr + _pos); \
} \
} while (0)
/* vectorization
* older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax,
@@ -166,9 +192,9 @@
#endif
#if __has_builtin(__builtin_unreachable) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)))
# define ZSTD_UNREACHABLE { assert(0), __builtin_unreachable(); }
# define ZSTD_UNREACHABLE do { assert(0), __builtin_unreachable(); } while (0)
#else
# define ZSTD_UNREACHABLE { assert(0); }
# define ZSTD_UNREACHABLE do { assert(0); } while (0)
#endif
/* disable warnings */
@@ -181,30 +207,21 @@
# pragma warning(disable : 4324) /* disable: C4324: padded structure */
#endif
/*Like DYNAMIC_BMI2 but for compile time determination of BMI2 support*/
#ifndef STATIC_BMI2
# if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86))
# ifdef __AVX2__ //MSVC does not have a BMI2 specific flag, but every CPU that supports AVX2 also supports BMI2
# define STATIC_BMI2 1
# endif
# elif defined(__BMI2__) && defined(__x86_64__) && defined(__GNUC__)
# define STATIC_BMI2 1
# endif
#endif
#ifndef STATIC_BMI2
#define STATIC_BMI2 0
#endif
/* compile time determination of SIMD support */
#if !defined(ZSTD_NO_INTRINSICS)
# if defined(__SSE2__) || defined(_M_AMD64) || (defined (_M_IX86) && defined(_M_IX86_FP) && (_M_IX86_FP >= 2))
# if defined(__AVX2__)
# define ZSTD_ARCH_X86_AVX2
# endif
# if defined(__SSE2__) || defined(_M_X64) || (defined (_M_IX86) && defined(_M_IX86_FP) && (_M_IX86_FP >= 2))
# define ZSTD_ARCH_X86_SSE2
# endif
# if defined(__ARM_NEON) || defined(_M_ARM64)
# define ZSTD_ARCH_ARM_NEON
# endif
#
# if defined(ZSTD_ARCH_X86_AVX2)
# include <immintrin.h>
# endif
# if defined(ZSTD_ARCH_X86_SSE2)
# include <emmintrin.h>
# elif defined(ZSTD_ARCH_ARM_NEON)
@@ -249,9 +266,15 @@
#endif
/*-**************************************************************
* Alignment check
* Alignment
*****************************************************************/
/* @return 1 if @u is a 2^n value, 0 otherwise
* useful to check a value is valid for alignment restrictions */
MEM_STATIC int ZSTD_isPower2(size_t u) {
return (u & (u-1)) == 0;
}
/* this test was initially positioned in mem.h,
* but this file is removed (or replaced) for linux kernel
* so it's now hosted in compiler.h,
@@ -281,10 +304,93 @@
# endif
#endif /* ZSTD_ALIGNOF */
#ifndef ZSTD_ALIGNED
/* C90-compatible alignment macro (GCC/Clang). Adjust for other compilers if needed. */
# if defined(__GNUC__) || defined(__clang__)
# define ZSTD_ALIGNED(a) __attribute__((aligned(a)))
# elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */
# define ZSTD_ALIGNED(a) _Alignas(a)
#elif defined(_MSC_VER)
# define ZSTD_ALIGNED(n) __declspec(align(n))
# else
/* this compiler will require its own alignment instruction */
# define ZSTD_ALIGNED(...)
# endif
#endif /* ZSTD_ALIGNED */
/*-**************************************************************
* Sanitizer
*****************************************************************/
/**
* Zstd relies on pointer overflow in its decompressor.
* We add this attribute to functions that rely on pointer overflow.
*/
#ifndef ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
# if __has_attribute(no_sanitize)
# if !defined(__clang__) && defined(__GNUC__) && __GNUC__ < 8
/* gcc < 8 only has signed-integer-overlow which triggers on pointer overflow */
# define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("signed-integer-overflow")))
# else
/* older versions of clang [3.7, 5.0) will warn that pointer-overflow is ignored. */
# define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR __attribute__((no_sanitize("pointer-overflow")))
# endif
# else
# define ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
# endif
#endif
/**
* Helper function to perform a wrapped pointer difference without triggering
* UBSAN.
*
* @returns lhs - rhs with wrapping
*/
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
ptrdiff_t ZSTD_wrappedPtrDiff(unsigned char const* lhs, unsigned char const* rhs)
{
return lhs - rhs;
}
/**
* Helper function to perform a wrapped pointer add without triggering UBSAN.
*
* @return ptr + add with wrapping
*/
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
unsigned char const* ZSTD_wrappedPtrAdd(unsigned char const* ptr, ptrdiff_t add)
{
return ptr + add;
}
/**
* Helper function to perform a wrapped pointer subtraction without triggering
* UBSAN.
*
* @return ptr - sub with wrapping
*/
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
unsigned char const* ZSTD_wrappedPtrSub(unsigned char const* ptr, ptrdiff_t sub)
{
return ptr - sub;
}
/**
* Helper function to add to a pointer that works around C's undefined behavior
* of adding 0 to NULL.
*
* @returns `ptr + add` except it defines `NULL + 0 == NULL`.
*/
MEM_STATIC
unsigned char* ZSTD_maybeNullPtrAdd(unsigned char* ptr, ptrdiff_t add)
{
return add > 0 ? ptr + add : ptr;
}
/* Issue #3240 reports an ASAN failure on an llvm-mingw build. Out of an
* abundance of caution, disable our custom poisoning on mingw. */
#ifdef __MINGW32__
Utilities/cmzstd/lib/common/cpu.h
+36
View File
@@ -35,6 +35,7 @@ MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
U32 f7b = 0;
U32 f7c = 0;
#if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86))
#if !defined(_M_X64) || !defined(__clang__) || __clang_major__ >= 16
int reg[4];
__cpuid((int*)reg, 0);
{
@@ -50,6 +51,41 @@ MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
f7c = (U32)reg[2];
}
}
#else
/* Clang compiler has a bug (fixed in https://reviews.llvm.org/D101338) in
* which the `__cpuid` intrinsic does not save and restore `rbx` as it needs
* to due to being a reserved register. So in that case, do the `cpuid`
* ourselves. Clang supports inline assembly anyway.
*/
U32 n;
__asm__(
"pushq %%rbx\n\t"
"cpuid\n\t"
"popq %%rbx\n\t"
: "=a"(n)
: "a"(0)
: "rcx", "rdx");
if (n >= 1) {
U32 f1a;
__asm__(
"pushq %%rbx\n\t"
"cpuid\n\t"
"popq %%rbx\n\t"
: "=a"(f1a), "=c"(f1c), "=d"(f1d)
: "a"(1)
:);
}
if (n >= 7) {
__asm__(
"pushq %%rbx\n\t"
"cpuid\n\t"
"movq %%rbx, %%rax\n\t"
"popq %%rbx"
: "=a"(f7b), "=c"(f7c)
: "a"(7), "c"(0)
: "rdx");
}
#endif
#elif defined(__i386__) && defined(__PIC__) && !defined(__clang__) && defined(__GNUC__)
/* The following block like the normal cpuid branch below, but gcc
* reserves ebx for use of its pic register so we must specially
Utilities/cmzstd/lib/common/debug.c
+6
View File
@@ -21,4 +21,10 @@
#include "debug.h"
#if !defined(ZSTD_LINUX_KERNEL) || (DEBUGLEVEL>=2)
/* We only use this when DEBUGLEVEL>=2, but we get -Werror=pedantic errors if a
* translation unit is empty. So remove this from Linux kernel builds, but
* otherwise just leave it in.
*/
int g_debuglevel = DEBUGLEVEL;
#endif
Utilities/cmzstd/lib/common/debug.h
+20 -20
View File
@@ -32,10 +32,6 @@
#ifndef DEBUG_H_12987983217
#define DEBUG_H_12987983217
#if defined (__cplusplus)
extern "C" {
#endif
/* static assert is triggered at compile time, leaving no runtime artefact.
* static assert only works with compile-time constants.
@@ -85,23 +81,27 @@ extern int g_debuglevel; /* the variable is only declared,
It's useful when enabling very verbose levels
on selective conditions (such as position in src) */
# define RAWLOG(l, ...) { \
if (l<=g_debuglevel) { \
ZSTD_DEBUG_PRINT(__VA_ARGS__); \
} }
# define DEBUGLOG(l, ...) { \
if (l<=g_debuglevel) { \
ZSTD_DEBUG_PRINT(__FILE__ ": " __VA_ARGS__); \
ZSTD_DEBUG_PRINT(" \n"); \
} }
# define RAWLOG(l, ...) \
do { \
if (l<=g_debuglevel) { \
ZSTD_DEBUG_PRINT(__VA_ARGS__); \
} \
} while (0)
#define STRINGIFY(x) #x
#define TOSTRING(x) STRINGIFY(x)
#define LINE_AS_STRING TOSTRING(__LINE__)
# define DEBUGLOG(l, ...) \
do { \
if (l<=g_debuglevel) { \
ZSTD_DEBUG_PRINT(__FILE__ ":" LINE_AS_STRING ": " __VA_ARGS__); \
ZSTD_DEBUG_PRINT(" \n"); \
} \
} while (0)
#else
# define RAWLOG(l, ...) {} /* disabled */
# define DEBUGLOG(l, ...) {} /* disabled */
#endif
#if defined (__cplusplus)
}
# define RAWLOG(l, ...) do { } while (0) /* disabled */
# define DEBUGLOG(l, ...) do { } while (0) /* disabled */
#endif
#endif /* DEBUG_H_12987983217 */
Utilities/cmzstd/lib/common/error_private.c
+1
View File
@@ -40,6 +40,7 @@ const char* ERR_getErrorString(ERR_enum code)
case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported";
case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large";
case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small";
case PREFIX(cannotProduce_uncompressedBlock): return "This mode cannot generate an uncompressed block";
case PREFIX(stabilityCondition_notRespected): return "pledged buffer stability condition is not respected";
case PREFIX(dictionary_corrupted): return "Dictionary is corrupted";
case PREFIX(dictionary_wrong): return "Dictionary mismatch";
Utilities/cmzstd/lib/common/error_private.h
+45 -46
View File
@@ -13,11 +13,6 @@
#ifndef ERROR_H_MODULE
#define ERROR_H_MODULE
#if defined (__cplusplus)
extern "C" {
#endif
/* ****************************************
* Dependencies
******************************************/
@@ -26,7 +21,6 @@ extern "C" {
#include "debug.h"
#include "zstd_deps.h" /* size_t */
/* ****************************************
* Compiler-specific
******************************************/
@@ -60,8 +54,13 @@ ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }
ERR_STATIC ERR_enum ERR_getErrorCode(size_t code) { if (!ERR_isError(code)) return (ERR_enum)0; return (ERR_enum) (0-code); }
/* check and forward error code */
#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e
#define CHECK_F(f) { CHECK_V_F(_var_err__, f); }
#define CHECK_V_F(e, f) \
size_t const e = f; \
do { \
if (ERR_isError(e)) \
return e; \
} while (0)
#define CHECK_F(f) do { CHECK_V_F(_var_err__, f); } while (0)
/*-****************************************
@@ -95,10 +94,12 @@ void _force_has_format_string(const char *format, ...) {
* We want to force this function invocation to be syntactically correct, but
* we don't want to force runtime evaluation of its arguments.
*/
#define _FORCE_HAS_FORMAT_STRING(...) \
if (0) { \
_force_has_format_string(__VA_ARGS__); \
}
#define _FORCE_HAS_FORMAT_STRING(...) \
do { \
if (0) { \
_force_has_format_string(__VA_ARGS__); \
} \
} while (0)
#define ERR_QUOTE(str) #str
@@ -109,51 +110,49 @@ void _force_has_format_string(const char *format, ...) {
* In order to do that (particularly, printing the conditional that failed),
* this can't just wrap RETURN_ERROR().
*/
#define RETURN_ERROR_IF(cond, err, ...) \
if (cond) { \
RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", \
__FILE__, __LINE__, ERR_QUOTE(cond), ERR_QUOTE(ERROR(err))); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return ERROR(err); \
}
#define RETURN_ERROR_IF(cond, err, ...) \
do { \
if (cond) { \
RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", \
__FILE__, __LINE__, ERR_QUOTE(cond), ERR_QUOTE(ERROR(err))); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return ERROR(err); \
} \
} while (0)
/**
* Unconditionally return the specified error.
*
* In debug modes, prints additional information.
*/
#define RETURN_ERROR(err, ...) \
do { \
RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", \
__FILE__, __LINE__, ERR_QUOTE(ERROR(err))); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return ERROR(err); \
} while(0);
#define RETURN_ERROR(err, ...) \
do { \
RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", \
__FILE__, __LINE__, ERR_QUOTE(ERROR(err))); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return ERROR(err); \
} while(0)
/**
* If the provided expression evaluates to an error code, returns that error code.
*
* In debug modes, prints additional information.
*/
#define FORWARD_IF_ERROR(err, ...) \
do { \
size_t const err_code = (err); \
if (ERR_isError(err_code)) { \
RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", \
__FILE__, __LINE__, ERR_QUOTE(err), ERR_getErrorName(err_code)); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return err_code; \
} \
} while(0);
#if defined (__cplusplus)
}
#endif
#define FORWARD_IF_ERROR(err, ...) \
do { \
size_t const err_code = (err); \
if (ERR_isError(err_code)) { \
RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", \
__FILE__, __LINE__, ERR_QUOTE(err), ERR_getErrorName(err_code)); \
_FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
RAWLOG(3, ": " __VA_ARGS__); \
RAWLOG(3, "\n"); \
return err_code; \
} \
} while(0)
#endif /* ERROR_H_MODULE */
Utilities/cmzstd/lib/common/fse.h
+3 -17
View File
@@ -11,11 +11,6 @@
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
****************************************************************** */
#if defined (__cplusplus)
extern "C" {
#endif
#ifndef FSE_H
#define FSE_H
@@ -25,7 +20,6 @@ extern "C" {
******************************************/
#include "zstd_deps.h" /* size_t, ptrdiff_t */
/*-*****************************************
* FSE_PUBLIC_API : control library symbols visibility
******************************************/
@@ -229,13 +223,11 @@ If there is an error, the function will return an error code, which can be teste
#endif /* FSE_H */
#if defined(FSE_STATIC_LINKING_ONLY) && !defined(FSE_H_FSE_STATIC_LINKING_ONLY)
#define FSE_H_FSE_STATIC_LINKING_ONLY
/* *** Dependency *** */
#include "bitstream.h"
/* *****************************************
* Static allocation
*******************************************/
@@ -464,13 +456,13 @@ MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, un
FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
const U16* const stateTable = (const U16*)(statePtr->stateTable);
U32 const nbBitsOut = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
BIT_addBits(bitC, statePtr->value, nbBitsOut);
BIT_addBits(bitC, (BitContainerType)statePtr->value, nbBitsOut);
statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
}
MEM_STATIC void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* statePtr)
{
BIT_addBits(bitC, statePtr->value, statePtr->stateLog);
BIT_addBits(bitC, (BitContainerType)statePtr->value, statePtr->stateLog);
BIT_flushBits(bitC);
}
@@ -630,10 +622,4 @@ MEM_STATIC unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr)
#define FSE_TABLESTEP(tableSize) (((tableSize)>>1) + ((tableSize)>>3) + 3)
#endif /* FSE_STATIC_LINKING_ONLY */
#if defined (__cplusplus)
}
#endif
Utilities/cmzstd/lib/common/fse_decompress.c
+21 -17
View File
@@ -22,8 +22,7 @@
#define FSE_STATIC_LINKING_ONLY
#include "fse.h"
#include "error_private.h"
#define ZSTD_DEPS_NEED_MALLOC
#include "zstd_deps.h"
#include "zstd_deps.h" /* ZSTD_memcpy */
#include "bits.h" /* ZSTD_highbit32 */
@@ -84,7 +83,7 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
symbolNext[s] = 1;
} else {
if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
symbolNext[s] = normalizedCounter[s];
symbolNext[s] = (U16)normalizedCounter[s];
} } }
ZSTD_memcpy(dt, &DTableH, sizeof(DTableH));
}
@@ -99,8 +98,7 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
* all symbols have counts <= 8. We ensure we have 8 bytes at the end of
* our buffer to handle the over-write.
*/
{
U64 const add = 0x0101010101010101ull;
{ U64 const add = 0x0101010101010101ull;
size_t pos = 0;
U64 sv = 0;
U32 s;
@@ -111,9 +109,8 @@ static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCo
for (i = 8; i < n; i += 8) {
MEM_write64(spread + pos + i, sv);
}
pos += n;
}
}
pos += (size_t)n;
} }
/* Now we spread those positions across the table.
* The benefit of doing it in two stages is that we avoid the
* variable size inner loop, which caused lots of branch misses.
@@ -193,6 +190,8 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_usingDTable_generic(
FSE_initDState(&state1, &bitD, dt);
FSE_initDState(&state2, &bitD, dt);
RETURN_ERROR_IF(BIT_reloadDStream(&bitD)==BIT_DStream_overflow, corruption_detected, "");
#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD)
/* 4 symbols per loop */
@@ -232,12 +231,12 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_usingDTable_generic(
break;
} }
return op-ostart;
assert(op >= ostart);
return (size_t)(op-ostart);
}
typedef struct {
short ncount[FSE_MAX_SYMBOL_VALUE + 1];
FSE_DTable dtable[1]; /* Dynamically sized */
} FSE_DecompressWksp;
@@ -252,13 +251,18 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
unsigned tableLog;
unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
FSE_DecompressWksp* const wksp = (FSE_DecompressWksp*)workSpace;
size_t const dtablePos = sizeof(FSE_DecompressWksp) / sizeof(FSE_DTable);
FSE_DTable* const dtable = (FSE_DTable*)workSpace + dtablePos;
DEBUG_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
FSE_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
if (wkspSize < sizeof(*wksp)) return ERROR(GENERIC);
/* correct offset to dtable depends on this property */
FSE_STATIC_ASSERT(sizeof(FSE_DecompressWksp) % sizeof(FSE_DTable) == 0);
/* normal FSE decoding mode */
{
size_t const NCountLength = FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
{ size_t const NCountLength =
FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
if (FSE_isError(NCountLength)) return NCountLength;
if (tableLog > maxLog) return ERROR(tableLog_tooLarge);
assert(NCountLength <= cSrcSize);
@@ -271,16 +275,16 @@ FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
workSpace = (BYTE*)workSpace + sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
wkspSize -= sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
CHECK_F( FSE_buildDTable_internal(wksp->dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
CHECK_F( FSE_buildDTable_internal(dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
{
const void* ptr = wksp->dtable;
const void* ptr = dtable;
const FSE_DTableHeader* DTableH = (const FSE_DTableHeader*)ptr;
const U32 fastMode = DTableH->fastMode;
/* select fast mode (static) */
if (fastMode) return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 1);
return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 0);
if (fastMode) return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, dtable, 1);
return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, dtable, 0);
}
}
Utilities/cmzstd/lib/common/huf.h
+14 -10
View File
@@ -12,10 +12,6 @@
* You may select, at your option, one of the above-listed licenses.
****************************************************************** */
#if defined (__cplusplus)
extern "C" {
#endif
#ifndef HUF_H_298734234
#define HUF_H_298734234
@@ -25,7 +21,6 @@ extern "C" {
#define FSE_STATIC_LINKING_ONLY
#include "fse.h"
/* *** Tool functions *** */
#define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */
size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */
@@ -197,9 +192,22 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
/** HUF_getNbBitsFromCTable() :
* Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX
* Note 1 : is not inlined, as HUF_CElt definition is private */
* Note 1 : If symbolValue > HUF_readCTableHeader(symbolTable).maxSymbolValue, returns 0
* Note 2 : is not inlined, as HUF_CElt definition is private
*/
U32 HUF_getNbBitsFromCTable(const HUF_CElt* symbolTable, U32 symbolValue);
typedef struct {
BYTE tableLog;
BYTE maxSymbolValue;
BYTE unused[sizeof(size_t) - 2];
} HUF_CTableHeader;
/** HUF_readCTableHeader() :
* @returns The header from the CTable specifying the tableLog and the maxSymbolValue.
*/
HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable);
/*
* HUF_decompress() does the following:
* 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics
@@ -267,7 +275,3 @@ size_t HUF_readDTableX2_wksp(HUF_DTable* DTable, const void* src, size_t srcSize
#endif
#endif /* HUF_H_298734234 */
#if defined (__cplusplus)
}
#endif
Utilities/cmzstd/lib/common/mem.h
+7 -20
View File
@@ -11,10 +11,6 @@
#ifndef MEM_H_MODULE
#define MEM_H_MODULE
#if defined (__cplusplus)
extern "C" {
#endif
/*-****************************************
* Dependencies
******************************************/
@@ -30,15 +26,8 @@ extern "C" {
#if defined(_MSC_VER) /* Visual Studio */
# include <stdlib.h> /* _byteswap_ulong */
# include <intrin.h> /* _byteswap_* */
#endif
#if defined(__GNUC__)
# define MEM_STATIC static __inline __attribute__((unused))
#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
# define MEM_STATIC static inline
#elif defined(_MSC_VER)
# define MEM_STATIC static __inline
#else
# define MEM_STATIC static /* this version may generate warnings for unused static functions; disable the relevant warning */
#elif defined(__ICCARM__)
# include <intrinsics.h>
#endif
/*-**************************************************************
@@ -83,7 +72,6 @@ extern "C" {
typedef signed long long S64;
#endif
/*-**************************************************************
* Memory I/O API
*****************************************************************/
@@ -159,10 +147,12 @@ MEM_STATIC unsigned MEM_isLittleEndian(void)
return 1;
#elif defined(__clang__) && __BIG_ENDIAN__
return 0;
#elif defined(_MSC_VER) && (_M_AMD64 || _M_IX86)
#elif defined(_MSC_VER) && (_M_X64 || _M_IX86)
return 1;
#elif defined(__DMC__) && defined(_M_IX86)
return 1;
#elif defined(__IAR_SYSTEMS_ICC__) && __LITTLE_ENDIAN__
return 1;
#else
const union { U32 u; BYTE c[4]; } one = { 1 }; /* don't use static : performance detrimental */
return one.c[0];
@@ -255,6 +245,8 @@ MEM_STATIC U32 MEM_swap32(U32 in)
#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
|| (defined(__clang__) && __has_builtin(__builtin_bswap32))
return __builtin_bswap32(in);
#elif defined(__ICCARM__)
return __REV(in);
#else
return MEM_swap32_fallback(in);
#endif
@@ -427,9 +419,4 @@ MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val)
/* code only tested on 32 and 64 bits systems */
MEM_STATIC void MEM_check(void) { DEBUG_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }
#if defined (__cplusplus)
}
#endif
#endif /* MEM_H_MODULE */
Utilities/cmzstd/lib/common/pool.c
+1 -1
View File
@@ -223,7 +223,7 @@ static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)
{ ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
if (!threadPool) return 1;
/* replace existing thread pool */
ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(ZSTD_pthread_t));
ZSTD_customFree(ctx->threads, ctx->customMem);
ctx->threads = threadPool;
/* Initialize additional threads */
Utilities/cmzstd/lib/common/pool.h
+1 -10
View File
@@ -11,10 +11,6 @@
#ifndef POOL_H
#define POOL_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "zstd_deps.h"
#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_customMem */
@@ -47,7 +43,7 @@ void POOL_joinJobs(POOL_ctx* ctx);
/*! POOL_resize() :
* Expands or shrinks pool's number of threads.
* This is more efficient than releasing + creating a new context,
* since it tries to preserve and re-use existing threads.
* since it tries to preserve and reuse existing threads.
* `numThreads` must be at least 1.
* @return : 0 when resize was successful,
* !0 (typically 1) if there is an error.
@@ -82,9 +78,4 @@ void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque);
*/
int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque);
#if defined (__cplusplus)
}
#endif
#endif
Utilities/cmzstd/lib/common/portability_macros.h
+24 -9
View File
@@ -68,30 +68,45 @@
/* Mark the internal assembly functions as hidden */
#ifdef __ELF__
# define ZSTD_HIDE_ASM_FUNCTION(func) .hidden func
#elif defined(__APPLE__)
# define ZSTD_HIDE_ASM_FUNCTION(func) .private_extern func
#else
# define ZSTD_HIDE_ASM_FUNCTION(func)
#endif
/* Compile time determination of BMI2 support */
#ifndef STATIC_BMI2
# if defined(__BMI2__)
# define STATIC_BMI2 1
# elif defined(_MSC_VER) && defined(__AVX2__)
# define STATIC_BMI2 1 /* MSVC does not have a BMI2 specific flag, but every CPU that supports AVX2 also supports BMI2 */
# endif
#endif
#ifndef STATIC_BMI2
# define STATIC_BMI2 0
#endif
/* Enable runtime BMI2 dispatch based on the CPU.
* Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.
*/
#ifndef DYNAMIC_BMI2
#if ((defined(__clang__) && __has_attribute(__target__)) \
# if ((defined(__clang__) && __has_attribute(__target__)) \
|| (defined(__GNUC__) \
&& (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \
&& (defined(__x86_64__) || defined(_M_X64)) \
&& (defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64)) \
&& !defined(__BMI2__)
# define DYNAMIC_BMI2 1
#else
# define DYNAMIC_BMI2 0
#endif
# define DYNAMIC_BMI2 1
# else
# define DYNAMIC_BMI2 0
# endif
#endif
/**
* Only enable assembly for GNUC compatible compilers,
* Only enable assembly for GNU C compatible compilers,
* because other platforms may not support GAS assembly syntax.
*
* Only enable assembly for Linux / MacOS, other platforms may
* Only enable assembly for Linux / MacOS / Win32, other platforms may
* work, but they haven't been tested. This could likely be
* extended to BSD systems.
*
@@ -99,7 +114,7 @@
* 100% of code to be instrumented to work.
*/
#if defined(__GNUC__)
# if defined(__linux__) || defined(__linux) || defined(__APPLE__)
# if defined(__linux__) || defined(__linux) || defined(__APPLE__) || defined(_WIN32)
# if ZSTD_MEMORY_SANITIZER
# define ZSTD_ASM_SUPPORTED 0
# elif ZSTD_DATAFLOW_SANITIZER
Utilities/cmzstd/lib/common/threading.c
+8 -2
View File
@@ -73,10 +73,12 @@ int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
ZSTD_thread_params_t thread_param;
(void)unused;
if (thread==NULL) return -1;
*thread = NULL;
thread_param.start_routine = start_routine;
thread_param.arg = arg;
thread_param.initialized = 0;
*thread = NULL;
/* Setup thread initialization synchronization */
if(ZSTD_pthread_cond_init(&thread_param.initialized_cond, NULL)) {
@@ -91,7 +93,7 @@ int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
/* Spawn thread */
*thread = (HANDLE)_beginthreadex(NULL, 0, worker, &thread_param, 0, NULL);
if (!thread) {
if (*thread==NULL) {
ZSTD_pthread_mutex_destroy(&thread_param.initialized_mutex);
ZSTD_pthread_cond_destroy(&thread_param.initialized_cond);
return errno;
@@ -137,6 +139,7 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread)
int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr)
{
assert(mutex != NULL);
*mutex = (pthread_mutex_t*)ZSTD_malloc(sizeof(pthread_mutex_t));
if (!*mutex)
return 1;
@@ -145,6 +148,7 @@ int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t con
int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
{
assert(mutex != NULL);
if (!*mutex)
return 0;
{
@@ -156,6 +160,7 @@ int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr)
{
assert(cond != NULL);
*cond = (pthread_cond_t*)ZSTD_malloc(sizeof(pthread_cond_t));
if (!*cond)
return 1;
@@ -164,6 +169,7 @@ int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const*
int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond)
{
assert(cond != NULL);
if (!*cond)
return 0;
{
Utilities/cmzstd/lib/common/threading.h
-8
View File
@@ -16,10 +16,6 @@
#include "debug.h"
#if defined (__cplusplus)
extern "C" {
#endif
#if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
/**
@@ -72,7 +68,6 @@ int ZSTD_pthread_join(ZSTD_pthread_t thread);
* add here more wrappers as required
*/
#elif defined(ZSTD_MULTITHREAD) /* posix assumed ; need a better detection method */
/* === POSIX Systems === */
# include <pthread.h>
@@ -143,8 +138,5 @@ typedef int ZSTD_pthread_cond_t;
#endif /* ZSTD_MULTITHREAD */
#if defined (__cplusplus)
}
#endif
#endif /* THREADING_H_938743 */
Utilities/cmzstd/lib/common/xxhash.c
+5 -11
View File
@@ -1,24 +1,18 @@
/*
* xxHash - Fast Hash algorithm
* Copyright (c) Meta Platforms, Inc. and affiliates.
*
* You can contact the author at :
* - xxHash homepage: https://cyan4973.github.io/xxHash/
* - xxHash source repository : https://github.com/Cyan4973/xxHash
* xxHash - Extremely Fast Hash algorithm
* Copyright (c) Yann Collet - Meta Platforms, Inc
*
* 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.
*/
*/
/*
* xxhash.c instantiates functions defined in xxhash.h
*/
#define XXH_STATIC_LINKING_ONLY /* access advanced declarations */
#define XXH_IMPLEMENTATION /* access definitions */
#define XXH_STATIC_LINKING_ONLY /* access advanced declarations */
#define XXH_IMPLEMENTATION /* access definitions */
#include "xxhash.h"
Utilities/cmzstd/lib/common/xxhash.h
+2413 -1005
View File
File diff suppressed because it is too large. Load diff
Utilities/cmzstd/lib/common/zstd_deps.h
+12
View File
@@ -24,6 +24,18 @@
#ifndef ZSTD_DEPS_COMMON
#define ZSTD_DEPS_COMMON
/* Even though we use qsort_r only for the dictionary builder, the macro
* _GNU_SOURCE has to be declared *before* the inclusion of any standard
* header and the script 'combine.sh' combines the whole zstd source code
* in a single file.
*/
#if defined(__linux) || defined(__linux__) || defined(linux) || defined(__gnu_linux__) || \
defined(__CYGWIN__) || defined(__MSYS__)
#if !defined(_GNU_SOURCE) && !defined(__ANDROID__) /* NDK doesn't ship qsort_r(). */
#define _GNU_SOURCE
#endif
#endif
#include <limits.h>
#include <stddef.h>
#include <string.h>
Utilities/cmzstd/lib/common/zstd_internal.h
+6 -74
View File
@@ -39,10 +39,6 @@
# define ZSTD_TRACE 0
#endif
#if defined (__cplusplus)
extern "C" {
#endif
/* ---- static assert (debug) --- */
#define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)
#define ZSTD_isError ERR_isError /* for inlining */
@@ -95,7 +91,7 @@ typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;
#define MIN_CBLOCK_SIZE (1 /*litCSize*/ + 1 /* RLE or RAW */) /* for a non-null block */
#define MIN_LITERALS_FOR_4_STREAMS 6
typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingType_e;
typedef enum { set_basic, set_rle, set_compressed, set_repeat } SymbolEncodingType_e;
#define LONGNBSEQ 0x7F00
@@ -178,7 +174,7 @@ static void ZSTD_copy8(void* dst, const void* src) {
ZSTD_memcpy(dst, src, 8);
#endif
}
#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
#define COPY8(d,s) do { ZSTD_copy8(d,s); d+=8; s+=8; } while (0)
/* Need to use memmove here since the literal buffer can now be located within
the dst buffer. In circumstances where the op "catches up" to where the
@@ -198,7 +194,7 @@ static void ZSTD_copy16(void* dst, const void* src) {
ZSTD_memcpy(dst, copy16_buf, 16);
#endif
}
#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
#define COPY16(d,s) do { ZSTD_copy16(d,s); d+=16; s+=16; } while (0)
#define WILDCOPY_OVERLENGTH 32
#define WILDCOPY_VECLEN 16
@@ -227,7 +223,7 @@ void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e
if (ovtype == ZSTD_overlap_src_before_dst && diff < WILDCOPY_VECLEN) {
/* Handle short offset copies. */
do {
COPY8(op, ip)
COPY8(op, ip);
} while (op < oend);
} else {
assert(diff >= WILDCOPY_VECLEN || diff <= -WILDCOPY_VECLEN);
@@ -278,62 +274,6 @@ typedef enum {
/*-*******************************************
* Private declarations
*********************************************/
typedef struct seqDef_s {
U32 offBase; /* offBase == Offset + ZSTD_REP_NUM, or repcode 1,2,3 */
U16 litLength;
U16 mlBase; /* mlBase == matchLength - MINMATCH */
} seqDef;
/* Controls whether seqStore has a single "long" litLength or matchLength. See seqStore_t. */
typedef enum {
ZSTD_llt_none = 0, /* no longLengthType */
ZSTD_llt_literalLength = 1, /* represents a long literal */
ZSTD_llt_matchLength = 2 /* represents a long match */
} ZSTD_longLengthType_e;
typedef struct {
seqDef* sequencesStart;
seqDef* sequences; /* ptr to end of sequences */
BYTE* litStart;
BYTE* lit; /* ptr to end of literals */
BYTE* llCode;
BYTE* mlCode;
BYTE* ofCode;
size_t maxNbSeq;
size_t maxNbLit;
/* longLengthPos and longLengthType to allow us to represent either a single litLength or matchLength
* in the seqStore that has a value larger than U16 (if it exists). To do so, we increment
* the existing value of the litLength or matchLength by 0x10000.
*/
ZSTD_longLengthType_e longLengthType;
U32 longLengthPos; /* Index of the sequence to apply long length modification to */
} seqStore_t;
typedef struct {
U32 litLength;
U32 matchLength;
} ZSTD_sequenceLength;
/**
* Returns the ZSTD_sequenceLength for the given sequences. It handles the decoding of long sequences
* indicated by longLengthPos and longLengthType, and adds MINMATCH back to matchLength.
*/
MEM_STATIC ZSTD_sequenceLength ZSTD_getSequenceLength(seqStore_t const* seqStore, seqDef const* seq)
{
ZSTD_sequenceLength seqLen;
seqLen.litLength = seq->litLength;
seqLen.matchLength = seq->mlBase + MINMATCH;
if (seqStore->longLengthPos == (U32)(seq - seqStore->sequencesStart)) {
if (seqStore->longLengthType == ZSTD_llt_literalLength) {
seqLen.litLength += 0x10000;
}
if (seqStore->longLengthType == ZSTD_llt_matchLength) {
seqLen.matchLength += 0x10000;
}
}
return seqLen;
}
/**
* Contains the compressed frame size and an upper-bound for the decompressed frame size.
@@ -347,10 +287,6 @@ typedef struct {
unsigned long long decompressedBound;
} ZSTD_frameSizeInfo; /* decompress & legacy */
const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx); /* compress & dictBuilder */
int ZSTD_seqToCodes(const seqStore_t* seqStorePtr); /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
/* ZSTD_invalidateRepCodes() :
* ensures next compression will not use repcodes from previous block.
* Note : only works with regular variant;
@@ -366,13 +302,13 @@ typedef struct {
/*! ZSTD_getcBlockSize() :
* Provides the size of compressed block from block header `src` */
/* Used by: decompress, fullbench (does not get its definition from here) */
/* Used by: decompress, fullbench */
size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
blockProperties_t* bpPtr);
/*! ZSTD_decodeSeqHeaders() :
* decode sequence header from src */
/* Used by: decompress, fullbench (does not get its definition from here) */
/* Used by: zstd_decompress_block, fullbench */
size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
const void* src, size_t srcSize);
@@ -385,8 +321,4 @@ MEM_STATIC int ZSTD_cpuSupportsBmi2(void)
return ZSTD_cpuid_bmi1(cpuid) && ZSTD_cpuid_bmi2(cpuid);
}
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_CCOMMON_H_MODULE */
Utilities/cmzstd/lib/common/zstd_trace.h
+5 -12
View File
@@ -11,23 +11,20 @@
#ifndef ZSTD_TRACE_H
#define ZSTD_TRACE_H
#if defined (__cplusplus)
extern "C" {
#endif
#include <stddef.h>
/* weak symbol support
* For now, enable conservatively:
* - Only GNUC
* - Only ELF
* - Only x86-64, i386 and aarch64
* - Only x86-64, i386, aarch64 and risc-v.
* Also, explicitly disable on platforms known not to work so they aren't
* forgotten in the future.
*/
#if !defined(ZSTD_HAVE_WEAK_SYMBOLS) && \
defined(__GNUC__) && defined(__ELF__) && \
(defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86) || defined(__aarch64__)) && \
(defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || \
defined(_M_IX86) || defined(__aarch64__) || defined(__riscv)) && \
!defined(__APPLE__) && !defined(_WIN32) && !defined(__MINGW32__) && \
!defined(__CYGWIN__) && !defined(_AIX)
# define ZSTD_HAVE_WEAK_SYMBOLS 1
@@ -64,7 +61,7 @@ typedef struct {
/**
* Non-zero if streaming (de)compression is used.
*/
unsigned streaming;
int streaming;
/**
* The dictionary ID.
*/
@@ -73,7 +70,7 @@ typedef struct {
* Is the dictionary cold?
* Only set on decompression.
*/
unsigned dictionaryIsCold;
int dictionaryIsCold;
/**
* The dictionary size or zero if no dictionary.
*/
@@ -156,8 +153,4 @@ ZSTD_WEAK_ATTR void ZSTD_trace_decompress_end(
#endif /* ZSTD_TRACE */
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_TRACE_H */
Utilities/cmzstd/lib/compress/fse_compress.c
+8 -7
View File
@@ -25,7 +25,7 @@
#include "../common/error_private.h"
#define ZSTD_DEPS_NEED_MALLOC
#define ZSTD_DEPS_NEED_MATH64
#include "../common/zstd_deps.h" /* ZSTD_malloc, ZSTD_free, ZSTD_memcpy, ZSTD_memset */
#include "../common/zstd_deps.h" /* ZSTD_memset */
#include "../common/bits.h" /* ZSTD_highbit32 */
@@ -225,8 +225,8 @@ size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog)
size_t const maxHeaderSize = (((maxSymbolValue+1) * tableLog
+ 4 /* bitCount initialized at 4 */
+ 2 /* first two symbols may use one additional bit each */) / 8)
+ 1 /* round up to whole nb bytes */
+ 2 /* additional two bytes for bitstream flush */;
+ 1 /* round up to whole nb bytes */
+ 2 /* additional two bytes for bitstream flush */;
return maxSymbolValue ? maxHeaderSize : FSE_NCOUNTBOUND; /* maxSymbolValue==0 ? use default */
}
@@ -255,7 +255,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
/* Init */
remaining = tableSize+1; /* +1 for extra accuracy */
threshold = tableSize;
nbBits = tableLog+1;
nbBits = (int)tableLog+1;
while ((symbol < alphabetSize) && (remaining>1)) { /* stops at 1 */
if (previousIs0) {
@@ -274,7 +274,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
}
while (symbol >= start+3) {
start+=3;
bitStream += 3 << bitCount;
bitStream += 3U << bitCount;
bitCount += 2;
}
bitStream += (symbol-start) << bitCount;
@@ -294,7 +294,7 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
count++; /* +1 for extra accuracy */
if (count>=threshold)
count += max; /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */
bitStream += count << bitCount;
bitStream += (U32)count << bitCount;
bitCount += nbBits;
bitCount -= (count<max);
previousIs0 = (count==1);
@@ -322,7 +322,8 @@ FSE_writeNCount_generic (void* header, size_t headerBufferSize,
out[1] = (BYTE)(bitStream>>8);
out+= (bitCount+7) /8;
return (out-ostart);
assert(out >= ostart);
return (size_t)(out-ostart);
}
Utilities/cmzstd/lib/compress/hist.c
+10
View File
@@ -26,6 +26,16 @@ unsigned HIST_isError(size_t code) { return ERR_isError(code); }
/*-**************************************************************
* Histogram functions
****************************************************************/
void HIST_add(unsigned* count, const void* src, size_t srcSize)
{
const BYTE* ip = (const BYTE*)src;
const BYTE* const end = ip + srcSize;
while (ip<end) {
count[*ip++]++;
}
}
unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize)
{
Utilities/cmzstd/lib/compress/hist.h
+7
View File
@@ -73,3 +73,10 @@ size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
*/
unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize);
/*! HIST_add() :
* Lowest level: just add nb of occurrences of characters from @src into @count.
* @count is not reset. @count array is presumed large enough (i.e. 1 KB).
@ This function does not need any additional stack memory.
*/
void HIST_add(unsigned* count, const void* src, size_t srcSize);
Utilities/cmzstd/lib/compress/huf_compress.c
+54 -25
View File
@@ -220,6 +220,25 @@ static void HUF_setValue(HUF_CElt* elt, size_t value)
}
}
HUF_CTableHeader HUF_readCTableHeader(HUF_CElt const* ctable)
{
HUF_CTableHeader header;
ZSTD_memcpy(&header, ctable, sizeof(header));
return header;
}
static void HUF_writeCTableHeader(HUF_CElt* ctable, U32 tableLog, U32 maxSymbolValue)
{
HUF_CTableHeader header;
HUF_STATIC_ASSERT(sizeof(ctable[0]) == sizeof(header));
ZSTD_memset(&header, 0, sizeof(header));
assert(tableLog < 256);
header.tableLog = (BYTE)tableLog;
assert(maxSymbolValue < 256);
header.maxSymbolValue = (BYTE)maxSymbolValue;
ZSTD_memcpy(ctable, &header, sizeof(header));
}
typedef struct {
HUF_CompressWeightsWksp wksp;
BYTE bitsToWeight[HUF_TABLELOG_MAX + 1]; /* precomputed conversion table */
@@ -237,6 +256,9 @@ size_t HUF_writeCTable_wksp(void* dst, size_t maxDstSize,
HUF_STATIC_ASSERT(HUF_CTABLE_WORKSPACE_SIZE >= sizeof(HUF_WriteCTableWksp));
assert(HUF_readCTableHeader(CTable).maxSymbolValue == maxSymbolValue);
assert(HUF_readCTableHeader(CTable).tableLog == huffLog);
/* check conditions */
if (workspaceSize < sizeof(HUF_WriteCTableWksp)) return ERROR(GENERIC);
if (maxSymbolValue > HUF_SYMBOLVALUE_MAX) return ERROR(maxSymbolValue_tooLarge);
@@ -283,7 +305,9 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
if (tableLog > HUF_TABLELOG_MAX) return ERROR(tableLog_tooLarge);
if (nbSymbols > *maxSymbolValuePtr+1) return ERROR(maxSymbolValue_tooSmall);
CTable[0] = tableLog;
*maxSymbolValuePtr = nbSymbols - 1;
HUF_writeCTableHeader(CTable, tableLog, *maxSymbolValuePtr);
/* Prepare base value per rank */
{ U32 n, nextRankStart = 0;
@@ -315,7 +339,6 @@ size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void
{ U32 n; for (n=0; n<nbSymbols; n++) HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++); }
}
*maxSymbolValuePtr = nbSymbols - 1;
return readSize;
}
@@ -323,6 +346,8 @@ U32 HUF_getNbBitsFromCTable(HUF_CElt const* CTable, U32 symbolValue)
{
const HUF_CElt* const ct = CTable + 1;
assert(symbolValue <= HUF_SYMBOLVALUE_MAX);
if (symbolValue > HUF_readCTableHeader(CTable).maxSymbolValue)
return 0;
return (U32)HUF_getNbBits(ct[symbolValue]);
}
@@ -723,7 +748,8 @@ static void HUF_buildCTableFromTree(HUF_CElt* CTable, nodeElt const* huffNode, i
HUF_setNbBits(ct + huffNode[n].byte, huffNode[n].nbBits); /* push nbBits per symbol, symbol order */
for (n=0; n<alphabetSize; n++)
HUF_setValue(ct + n, valPerRank[HUF_getNbBits(ct[n])]++); /* assign value within rank, symbol order */
CTable[0] = maxNbBits;
HUF_writeCTableHeader(CTable, maxNbBits, maxSymbolValue);
}
size_t
@@ -776,13 +802,20 @@ size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count,
}
int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue) {
HUF_CElt const* ct = CTable + 1;
int bad = 0;
int s;
for (s = 0; s <= (int)maxSymbolValue; ++s) {
bad |= (count[s] != 0) & (HUF_getNbBits(ct[s]) == 0);
}
return !bad;
HUF_CTableHeader header = HUF_readCTableHeader(CTable);
HUF_CElt const* ct = CTable + 1;
int bad = 0;
int s;
assert(header.tableLog <= HUF_TABLELOG_ABSOLUTEMAX);
if (header.maxSymbolValue < maxSymbolValue)
return 0;
for (s = 0; s <= (int)maxSymbolValue; ++s) {
bad |= (count[s] != 0) & (HUF_getNbBits(ct[s]) == 0);
}
return !bad;
}
size_t HUF_compressBound(size_t size) { return HUF_COMPRESSBOUND(size); }
@@ -1024,17 +1057,17 @@ HUF_compress1X_usingCTable_internal_body(void* dst, size_t dstSize,
const void* src, size_t srcSize,
const HUF_CElt* CTable)
{
U32 const tableLog = (U32)CTable[0];
U32 const tableLog = HUF_readCTableHeader(CTable).tableLog;
HUF_CElt const* ct = CTable + 1;
const BYTE* ip = (const BYTE*) src;
BYTE* const ostart = (BYTE*)dst;
BYTE* const oend = ostart + dstSize;
BYTE* op = ostart;
HUF_CStream_t bitC;
/* init */
if (dstSize < 8) return 0; /* not enough space to compress */
{ size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op));
{ BYTE* op = ostart;
size_t const initErr = HUF_initCStream(&bitC, op, (size_t)(oend-op));
if (HUF_isError(initErr)) return 0; }
if (dstSize < HUF_tightCompressBound(srcSize, (size_t)tableLog) || tableLog > 11)
@@ -1255,7 +1288,7 @@ unsigned HUF_optimalTableLog(
{ BYTE* dst = (BYTE*)workSpace + sizeof(HUF_WriteCTableWksp);
size_t dstSize = wkspSize - sizeof(HUF_WriteCTableWksp);
size_t maxBits, hSize, newSize;
size_t hSize, newSize;
const unsigned symbolCardinality = HUF_cardinality(count, maxSymbolValue);
const unsigned minTableLog = HUF_minTableLog(symbolCardinality);
size_t optSize = ((size_t) ~0) - 1;
@@ -1266,12 +1299,14 @@ unsigned HUF_optimalTableLog(
/* Search until size increases */
for (optLogGuess = minTableLog; optLogGuess <= maxTableLog; optLogGuess++) {
DEBUGLOG(7, "checking for huffLog=%u", optLogGuess);
maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize);
if (ERR_isError(maxBits)) continue;
if (maxBits < optLogGuess && optLogGuess > minTableLog) break;
{ size_t maxBits = HUF_buildCTable_wksp(table, count, maxSymbolValue, optLogGuess, workSpace, wkspSize);
if (ERR_isError(maxBits)) continue;
hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize);
if (maxBits < optLogGuess && optLogGuess > minTableLog) break;
hSize = HUF_writeCTable_wksp(dst, dstSize, table, maxSymbolValue, (U32)maxBits, workSpace, wkspSize);
}
if (ERR_isError(hSize)) continue;
@@ -1372,12 +1407,6 @@ HUF_compress_internal (void* dst, size_t dstSize,
huffLog = (U32)maxBits;
DEBUGLOG(6, "bit distribution completed (%zu symbols)", showCTableBits(table->CTable + 1, maxSymbolValue+1));
}
/* Zero unused symbols in CTable, so we can check it for validity */
{
size_t const ctableSize = HUF_CTABLE_SIZE_ST(maxSymbolValue);
size_t const unusedSize = sizeof(table->CTable) - ctableSize * sizeof(HUF_CElt);
ZSTD_memset(table->CTable + ctableSize, 0, unusedSize);
}
/* Write table description header */
{ CHECK_V_F(hSize, HUF_writeCTable_wksp(op, dstSize, table->CTable, maxSymbolValue, huffLog,
@@ -1420,7 +1449,7 @@ size_t HUF_compress1X_repeat (void* dst, size_t dstSize,
/* HUF_compress4X_repeat():
* compress input using 4 streams.
* consider skipping quickly
* re-use an existing huffman compression table */
* reuse an existing huffman compression table */
size_t HUF_compress4X_repeat (void* dst, size_t dstSize,
const void* src, size_t srcSize,
unsigned maxSymbolValue, unsigned huffLog,
Utilities/cmzstd/lib/compress/zstd_compress.c
+1325 -514
View File
File diff suppressed because it is too large. Load diff
Utilities/cmzstd/lib/compress/zstd_compress_internal.h
+257 -153
View File
@@ -24,10 +24,7 @@
# include "zstdmt_compress.h"
#endif
#include "../common/bits.h" /* ZSTD_highbit32, ZSTD_NbCommonBytes */
#if defined (__cplusplus)
extern "C" {
#endif
#include "zstd_preSplit.h" /* ZSTD_SLIPBLOCK_WORKSPACESIZE */
/*-*************************************
* Constants
@@ -39,7 +36,7 @@ extern "C" {
It's not a big deal though : candidate will just be sorted again.
Additionally, candidate position 1 will be lost.
But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss.
The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy.
The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table reuse with a different strategy.
This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
@@ -82,6 +79,70 @@ typedef struct {
ZSTD_fseCTables_t fse;
} ZSTD_entropyCTables_t;
/***********************************************
* Sequences *
***********************************************/
typedef struct SeqDef_s {
U32 offBase; /* offBase == Offset + ZSTD_REP_NUM, or repcode 1,2,3 */
U16 litLength;
U16 mlBase; /* mlBase == matchLength - MINMATCH */
} SeqDef;
/* Controls whether seqStore has a single "long" litLength or matchLength. See SeqStore_t. */
typedef enum {
ZSTD_llt_none = 0, /* no longLengthType */
ZSTD_llt_literalLength = 1, /* represents a long literal */
ZSTD_llt_matchLength = 2 /* represents a long match */
} ZSTD_longLengthType_e;
typedef struct {
SeqDef* sequencesStart;
SeqDef* sequences; /* ptr to end of sequences */
BYTE* litStart;
BYTE* lit; /* ptr to end of literals */
BYTE* llCode;
BYTE* mlCode;
BYTE* ofCode;
size_t maxNbSeq;
size_t maxNbLit;
/* longLengthPos and longLengthType to allow us to represent either a single litLength or matchLength
* in the seqStore that has a value larger than U16 (if it exists). To do so, we increment
* the existing value of the litLength or matchLength by 0x10000.
*/
ZSTD_longLengthType_e longLengthType;
U32 longLengthPos; /* Index of the sequence to apply long length modification to */
} SeqStore_t;
typedef struct {
U32 litLength;
U32 matchLength;
} ZSTD_SequenceLength;
/**
* Returns the ZSTD_SequenceLength for the given sequences. It handles the decoding of long sequences
* indicated by longLengthPos and longLengthType, and adds MINMATCH back to matchLength.
*/
MEM_STATIC ZSTD_SequenceLength ZSTD_getSequenceLength(SeqStore_t const* seqStore, SeqDef const* seq)
{
ZSTD_SequenceLength seqLen;
seqLen.litLength = seq->litLength;
seqLen.matchLength = seq->mlBase + MINMATCH;
if (seqStore->longLengthPos == (U32)(seq - seqStore->sequencesStart)) {
if (seqStore->longLengthType == ZSTD_llt_literalLength) {
seqLen.litLength += 0x10000;
}
if (seqStore->longLengthType == ZSTD_llt_matchLength) {
seqLen.matchLength += 0x10000;
}
}
return seqLen;
}
const SeqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx); /* compress & dictBuilder */
int ZSTD_seqToCodes(const SeqStore_t* seqStorePtr); /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
/***********************************************
* Entropy buffer statistics structs and funcs *
***********************************************/
@@ -91,7 +152,7 @@ typedef struct {
* hufDesSize refers to the size of huffman tree description in bytes.
* This metadata is populated in ZSTD_buildBlockEntropyStats_literals() */
typedef struct {
symbolEncodingType_e hType;
SymbolEncodingType_e hType;
BYTE hufDesBuffer[ZSTD_MAX_HUF_HEADER_SIZE];
size_t hufDesSize;
} ZSTD_hufCTablesMetadata_t;
@@ -102,9 +163,9 @@ typedef struct {
* fseTablesSize refers to the size of fse tables in bytes.
* This metadata is populated in ZSTD_buildBlockEntropyStats_sequences() */
typedef struct {
symbolEncodingType_e llType;
symbolEncodingType_e ofType;
symbolEncodingType_e mlType;
SymbolEncodingType_e llType;
SymbolEncodingType_e ofType;
SymbolEncodingType_e mlType;
BYTE fseTablesBuffer[ZSTD_MAX_FSE_HEADERS_SIZE];
size_t fseTablesSize;
size_t lastCountSize; /* This is to account for bug in 1.3.4. More detail in ZSTD_entropyCompressSeqStore_internal() */
@@ -119,7 +180,7 @@ typedef struct {
* Builds entropy for the block.
* @return : 0 on success or error code */
size_t ZSTD_buildBlockEntropyStats(
const seqStore_t* seqStorePtr,
const SeqStore_t* seqStorePtr,
const ZSTD_entropyCTables_t* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
const ZSTD_CCtx_params* cctxParams,
@@ -148,34 +209,29 @@ typedef struct {
stopped. posInSequence <= seq[pos].litLength + seq[pos].matchLength */
size_t size; /* The number of sequences. <= capacity. */
size_t capacity; /* The capacity starting from `seq` pointer */
} rawSeqStore_t;
} RawSeqStore_t;
UNUSED_ATTR static const RawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0};
typedef struct {
U32 idx; /* Index in array of ZSTD_Sequence */
U32 posInSequence; /* Position within sequence at idx */
size_t posInSrc; /* Number of bytes given by sequences provided so far */
} ZSTD_sequencePosition;
UNUSED_ATTR static const rawSeqStore_t kNullRawSeqStore = {NULL, 0, 0, 0, 0};
typedef struct {
int price;
U32 off;
U32 mlen;
U32 litlen;
U32 rep[ZSTD_REP_NUM];
int price; /* price from beginning of segment to this position */
U32 off; /* offset of previous match */
U32 mlen; /* length of previous match */
U32 litlen; /* nb of literals since previous match */
U32 rep[ZSTD_REP_NUM]; /* offset history after previous match */
} ZSTD_optimal_t;
typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e;
#define ZSTD_OPT_SIZE (ZSTD_OPT_NUM+3)
typedef struct {
/* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */
unsigned* litFreq; /* table of literals statistics, of size 256 */
unsigned* litLengthFreq; /* table of litLength statistics, of size (MaxLL+1) */
unsigned* matchLengthFreq; /* table of matchLength statistics, of size (MaxML+1) */
unsigned* offCodeFreq; /* table of offCode statistics, of size (MaxOff+1) */
ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_NUM+1 */
ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */
ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_SIZE */
ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_SIZE */
U32 litSum; /* nb of literals */
U32 litLengthSum; /* nb of litLength codes */
@@ -187,7 +243,7 @@ typedef struct {
U32 offCodeSumBasePrice; /* to compare to log2(offreq) */
ZSTD_OptPrice_e priceType; /* prices can be determined dynamically, or follow a pre-defined cost structure */
const ZSTD_entropyCTables_t* symbolCosts; /* pre-calculated dictionary statistics */
ZSTD_paramSwitch_e literalCompressionMode;
ZSTD_ParamSwitch_e literalCompressionMode;
} optState_t;
typedef struct {
@@ -209,11 +265,11 @@ typedef struct {
#define ZSTD_WINDOW_START_INDEX 2
typedef struct ZSTD_matchState_t ZSTD_matchState_t;
typedef struct ZSTD_MatchState_t ZSTD_MatchState_t;
#define ZSTD_ROW_HASH_CACHE_SIZE 8 /* Size of prefetching hash cache for row-based matchfinder */
struct ZSTD_matchState_t {
struct ZSTD_MatchState_t {
ZSTD_window_t window; /* State for window round buffer management */
U32 loadedDictEnd; /* index of end of dictionary, within context's referential.
* When loadedDictEnd != 0, a dictionary is in use, and still valid.
@@ -228,22 +284,22 @@ struct ZSTD_matchState_t {
U32 rowHashLog; /* For row-based matchfinder: Hashlog based on nb of rows in the hashTable.*/
BYTE* tagTable; /* For row-based matchFinder: A row-based table containing the hashes and head index. */
U32 hashCache[ZSTD_ROW_HASH_CACHE_SIZE]; /* For row-based matchFinder: a cache of hashes to improve speed */
U64 hashSalt; /* For row-based matchFinder: salts the hash for re-use of tag table */
U64 hashSalt; /* For row-based matchFinder: salts the hash for reuse of tag table */
U32 hashSaltEntropy; /* For row-based matchFinder: collects entropy for salt generation */
U32* hashTable;
U32* hashTable3;
U32* chainTable;
U32 forceNonContiguous; /* Non-zero if we should force non-contiguous load for the next window update. */
int forceNonContiguous; /* Non-zero if we should force non-contiguous load for the next window update. */
int dedicatedDictSearch; /* Indicates whether this matchState is using the
* dedicated dictionary search structure.
*/
optState_t opt; /* optimal parser state */
const ZSTD_matchState_t* dictMatchState;
const ZSTD_MatchState_t* dictMatchState;
ZSTD_compressionParameters cParams;
const rawSeqStore_t* ldmSeqStore;
const RawSeqStore_t* ldmSeqStore;
/* Controls prefetching in some dictMatchState matchfinders.
* This behavior is controlled from the cctx ms.
@@ -261,7 +317,7 @@ struct ZSTD_matchState_t {
typedef struct {
ZSTD_compressedBlockState_t* prevCBlock;
ZSTD_compressedBlockState_t* nextCBlock;
ZSTD_matchState_t matchState;
ZSTD_MatchState_t matchState;
} ZSTD_blockState_t;
typedef struct {
@@ -288,7 +344,7 @@ typedef struct {
} ldmState_t;
typedef struct {
ZSTD_paramSwitch_e enableLdm; /* ZSTD_ps_enable to enable LDM. ZSTD_ps_auto by default */
ZSTD_ParamSwitch_e enableLdm; /* ZSTD_ps_enable to enable LDM. ZSTD_ps_auto by default */
U32 hashLog; /* Log size of hashTable */
U32 bucketSizeLog; /* Log bucket size for collision resolution, at most 8 */
U32 minMatchLength; /* Minimum match length */
@@ -319,7 +375,7 @@ struct ZSTD_CCtx_params_s {
* There is no guarantee that hint is close to actual source size */
ZSTD_dictAttachPref_e attachDictPref;
ZSTD_paramSwitch_e literalCompressionMode;
ZSTD_ParamSwitch_e literalCompressionMode;
/* Multithreading: used to pass parameters to mtctx */
int nbWorkers;
@@ -338,14 +394,27 @@ struct ZSTD_CCtx_params_s {
ZSTD_bufferMode_e outBufferMode;
/* Sequence compression API */
ZSTD_sequenceFormat_e blockDelimiters;
ZSTD_SequenceFormat_e blockDelimiters;
int validateSequences;
/* Block splitting */
ZSTD_paramSwitch_e useBlockSplitter;
/* Block splitting
* @postBlockSplitter executes split analysis after sequences are produced,
* it's more accurate but consumes more resources.
* @preBlockSplitter_level splits before knowing sequences,
* it's more approximative but also cheaper.
* Valid @preBlockSplitter_level values range from 0 to 6 (included).
* 0 means auto, 1 means do not split,
* then levels are sorted in increasing cpu budget, from 2 (fastest) to 6 (slowest).
* Highest @preBlockSplitter_level combines well with @postBlockSplitter.
*/
ZSTD_ParamSwitch_e postBlockSplitter;
int preBlockSplitter_level;
/* Adjust the max block size*/
size_t maxBlockSize;
/* Param for deciding whether to use row-based matchfinder */
ZSTD_paramSwitch_e useRowMatchFinder;
ZSTD_ParamSwitch_e useRowMatchFinder;
/* Always load a dictionary in ext-dict mode (not prefix mode)? */
int deterministicRefPrefix;
@@ -354,26 +423,25 @@ struct ZSTD_CCtx_params_s {
ZSTD_customMem customMem;
/* Controls prefetching in some dictMatchState matchfinders */
ZSTD_paramSwitch_e prefetchCDictTables;
ZSTD_ParamSwitch_e prefetchCDictTables;
/* Controls whether zstd will fall back to an internal matchfinder
* if the external matchfinder returns an error code. */
int enableMatchFinderFallback;
/* Indicates whether an external matchfinder has been referenced.
* Users can't set this externally.
* It is set internally in ZSTD_registerSequenceProducer(). */
int useSequenceProducer;
/* Adjust the max block size*/
size_t maxBlockSize;
/* Parameters for the external sequence producer API.
* Users set these parameters through ZSTD_registerSequenceProducer().
* It is not possible to set these parameters individually through the public API. */
void* extSeqProdState;
ZSTD_sequenceProducer_F extSeqProdFunc;
/* Controls repcode search in external sequence parsing */
ZSTD_paramSwitch_e searchForExternalRepcodes;
ZSTD_ParamSwitch_e searchForExternalRepcodes;
}; /* typedef'd to ZSTD_CCtx_params within "zstd.h" */
#define COMPRESS_SEQUENCES_WORKSPACE_SIZE (sizeof(unsigned) * (MaxSeq + 2))
#define ENTROPY_WORKSPACE_SIZE (HUF_WORKSPACE_SIZE + COMPRESS_SEQUENCES_WORKSPACE_SIZE)
#define TMP_WORKSPACE_SIZE (MAX(ENTROPY_WORKSPACE_SIZE, ZSTD_SLIPBLOCK_WORKSPACESIZE))
/**
* Indicates whether this compression proceeds directly from user-provided
@@ -391,24 +459,16 @@ typedef enum {
*/
#define ZSTD_MAX_NB_BLOCK_SPLITS 196
typedef struct {
seqStore_t fullSeqStoreChunk;
seqStore_t firstHalfSeqStore;
seqStore_t secondHalfSeqStore;
seqStore_t currSeqStore;
seqStore_t nextSeqStore;
SeqStore_t fullSeqStoreChunk;
SeqStore_t firstHalfSeqStore;
SeqStore_t secondHalfSeqStore;
SeqStore_t currSeqStore;
SeqStore_t nextSeqStore;
U32 partitions[ZSTD_MAX_NB_BLOCK_SPLITS];
ZSTD_entropyCTablesMetadata_t entropyMetadata;
} ZSTD_blockSplitCtx;
/* Context for block-level external matchfinder API */
typedef struct {
void* mState;
ZSTD_sequenceProducer_F* mFinder;
ZSTD_Sequence* seqBuffer;
size_t seqBufferCapacity;
} ZSTD_externalMatchCtx;
struct ZSTD_CCtx_s {
ZSTD_compressionStage_e stage;
int cParamsChanged; /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */
@@ -420,7 +480,7 @@ struct ZSTD_CCtx_s {
size_t dictContentSize;
ZSTD_cwksp workspace; /* manages buffer for dynamic allocations */
size_t blockSize;
size_t blockSizeMax;
unsigned long long pledgedSrcSizePlusOne; /* this way, 0 (default) == unknown */
unsigned long long consumedSrcSize;
unsigned long long producedCSize;
@@ -432,13 +492,14 @@ struct ZSTD_CCtx_s {
int isFirstBlock;
int initialized;
seqStore_t seqStore; /* sequences storage ptrs */
SeqStore_t seqStore; /* sequences storage ptrs */
ldmState_t ldmState; /* long distance matching state */
rawSeq* ldmSequences; /* Storage for the ldm output sequences */
size_t maxNbLdmSequences;
rawSeqStore_t externSeqStore; /* Mutable reference to external sequences */
RawSeqStore_t externSeqStore; /* Mutable reference to external sequences */
ZSTD_blockState_t blockState;
U32* entropyWorkspace; /* entropy workspace of ENTROPY_WORKSPACE_SIZE bytes */
void* tmpWorkspace; /* used as substitute of stack space - must be aligned for S64 type */
size_t tmpWkspSize;
/* Whether we are streaming or not */
ZSTD_buffered_policy_e bufferedPolicy;
@@ -479,8 +540,9 @@ struct ZSTD_CCtx_s {
/* Workspace for block splitter */
ZSTD_blockSplitCtx blockSplitCtx;
/* Workspace for external matchfinder */
ZSTD_externalMatchCtx externalMatchCtx;
/* Buffer for output from external sequence producer */
ZSTD_Sequence* extSeqBuf;
size_t extSeqBufCapacity;
};
typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e;
@@ -511,12 +573,12 @@ typedef enum {
* behavior of taking both the source size and the dict size into account
* when selecting and adjusting parameters.
*/
} ZSTD_cParamMode_e;
} ZSTD_CParamMode_e;
typedef size_t (*ZSTD_blockCompressor) (
ZSTD_matchState_t* bs, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
typedef size_t (*ZSTD_BlockCompressor_f) (
ZSTD_MatchState_t* bs, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_paramSwitch_e rowMatchfinderMode, ZSTD_dictMode_e dictMode);
ZSTD_BlockCompressor_f ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_ParamSwitch_e rowMatchfinderMode, ZSTD_dictMode_e dictMode);
MEM_STATIC U32 ZSTD_LLcode(U32 litLength)
@@ -562,6 +624,25 @@ MEM_STATIC int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value)
return 1;
}
/* ZSTD_selectAddr:
* @return index >= lowLimit ? candidate : backup,
* tries to force branchless codegen. */
MEM_STATIC const BYTE*
ZSTD_selectAddr(U32 index, U32 lowLimit, const BYTE* candidate, const BYTE* backup)
{
#if defined(__GNUC__) && defined(__x86_64__)
__asm__ (
"cmp %1, %2\n"
"cmova %3, %0\n"
: "+r"(candidate)
: "r"(index), "r"(lowLimit), "r"(backup)
);
return candidate;
#else
return index >= lowLimit ? candidate : backup;
#endif
}
/* ZSTD_noCompressBlock() :
* Writes uncompressed block to dst buffer from given src.
* Returns the size of the block */
@@ -644,14 +725,55 @@ ZSTD_safecopyLiterals(BYTE* op, BYTE const* ip, BYTE const* const iend, BYTE con
#define OFFBASE_TO_OFFSET(o) (assert(OFFBASE_IS_OFFSET(o)), (o) - ZSTD_REP_NUM)
#define OFFBASE_TO_REPCODE(o) (assert(OFFBASE_IS_REPCODE(o)), (o)) /* returns ID 1,2,3 */
/*! ZSTD_storeSeqOnly() :
* Store a sequence (litlen, litPtr, offBase and matchLength) into SeqStore_t.
* Literals themselves are not copied, but @litPtr is updated.
* @offBase : Users should employ macros REPCODE_TO_OFFBASE() and OFFSET_TO_OFFBASE().
* @matchLength : must be >= MINMATCH
*/
HINT_INLINE UNUSED_ATTR void
ZSTD_storeSeqOnly(SeqStore_t* seqStorePtr,
size_t litLength,
U32 offBase,
size_t matchLength)
{
assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq);
/* literal Length */
assert(litLength <= ZSTD_BLOCKSIZE_MAX);
if (UNLIKELY(litLength>0xFFFF)) {
assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */
seqStorePtr->longLengthType = ZSTD_llt_literalLength;
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
}
seqStorePtr->sequences[0].litLength = (U16)litLength;
/* match offset */
seqStorePtr->sequences[0].offBase = offBase;
/* match Length */
assert(matchLength <= ZSTD_BLOCKSIZE_MAX);
assert(matchLength >= MINMATCH);
{ size_t const mlBase = matchLength - MINMATCH;
if (UNLIKELY(mlBase>0xFFFF)) {
assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */
seqStorePtr->longLengthType = ZSTD_llt_matchLength;
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
}
seqStorePtr->sequences[0].mlBase = (U16)mlBase;
}
seqStorePtr->sequences++;
}
/*! ZSTD_storeSeq() :
* Store a sequence (litlen, litPtr, offBase and matchLength) into seqStore_t.
* Store a sequence (litlen, litPtr, offBase and matchLength) into SeqStore_t.
* @offBase : Users should employ macros REPCODE_TO_OFFBASE() and OFFSET_TO_OFFBASE().
* @matchLength : must be >= MINMATCH
* Allowed to over-read literals up to litLimit.
*/
HINT_INLINE UNUSED_ATTR void
ZSTD_storeSeq(seqStore_t* seqStorePtr,
ZSTD_storeSeq(SeqStore_t* seqStorePtr,
size_t litLength, const BYTE* literals, const BYTE* litLimit,
U32 offBase,
size_t matchLength)
@@ -685,29 +807,7 @@ ZSTD_storeSeq(seqStore_t* seqStorePtr,
}
seqStorePtr->lit += litLength;
/* literal Length */
if (litLength>0xFFFF) {
assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */
seqStorePtr->longLengthType = ZSTD_llt_literalLength;
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
}
seqStorePtr->sequences[0].litLength = (U16)litLength;
/* match offset */
seqStorePtr->sequences[0].offBase = offBase;
/* match Length */
assert(matchLength >= MINMATCH);
{ size_t const mlBase = matchLength - MINMATCH;
if (mlBase>0xFFFF) {
assert(seqStorePtr->longLengthType == ZSTD_llt_none); /* there can only be a single long length */
seqStorePtr->longLengthType = ZSTD_llt_matchLength;
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
}
seqStorePtr->sequences[0].mlBase = (U16)mlBase;
}
seqStorePtr->sequences++;
ZSTD_storeSeqOnly(seqStorePtr, litLength, offBase, matchLength);
}
/* ZSTD_updateRep() :
@@ -736,12 +836,12 @@ ZSTD_updateRep(U32 rep[ZSTD_REP_NUM], U32 const offBase, U32 const ll0)
typedef struct repcodes_s {
U32 rep[3];
} repcodes_t;
} Repcodes_t;
MEM_STATIC repcodes_t
MEM_STATIC Repcodes_t
ZSTD_newRep(U32 const rep[ZSTD_REP_NUM], U32 const offBase, U32 const ll0)
{
repcodes_t newReps;
Repcodes_t newReps;
ZSTD_memcpy(&newReps, rep, sizeof(newReps));
ZSTD_updateRep(newReps.rep, offBase, ll0);
return newReps;
@@ -784,8 +884,8 @@ ZSTD_count_2segments(const BYTE* ip, const BYTE* match,
size_t const matchLength = ZSTD_count(ip, match, vEnd);
if (match + matchLength != mEnd) return matchLength;
DEBUGLOG(7, "ZSTD_count_2segments: found a 2-parts match (current length==%zu)", matchLength);
DEBUGLOG(7, "distance from match beginning to end dictionary = %zi", mEnd - match);
DEBUGLOG(7, "distance from current pos to end buffer = %zi", iEnd - ip);
DEBUGLOG(7, "distance from match beginning to end dictionary = %i", (int)(mEnd - match));
DEBUGLOG(7, "distance from current pos to end buffer = %i", (int)(iEnd - ip));
DEBUGLOG(7, "next byte : ip==%02X, istart==%02X", ip[matchLength], *iStart);
DEBUGLOG(7, "final match length = %zu", matchLength + ZSTD_count(ip+matchLength, iStart, iEnd));
return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd);
@@ -923,11 +1023,12 @@ MEM_STATIC U64 ZSTD_rollingHash_rotate(U64 hash, BYTE toRemove, BYTE toAdd, U64
/*-*************************************
* Round buffer management
***************************************/
#if (ZSTD_WINDOWLOG_MAX_64 > 31)
# error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX"
#endif
/* Max current allowed */
#define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX))
/* Max @current value allowed:
* In 32-bit mode: we want to avoid crossing the 2 GB limit,
* reducing risks of side effects in case of signed operations on indexes.
* In 64-bit mode: we want to ensure that adding the maximum job size (512 MB)
* doesn't overflow U32 index capacity (4 GB) */
#define ZSTD_CURRENT_MAX (MEM_64bits() ? 3500U MB : 2000U MB)
/* Maximum chunk size before overflow correction needs to be called again */
#define ZSTD_CHUNKSIZE_MAX \
( ((U32)-1) /* Maximum ending current index */ \
@@ -967,7 +1068,7 @@ MEM_STATIC U32 ZSTD_window_hasExtDict(ZSTD_window_t const window)
* Inspects the provided matchState and figures out what dictMode should be
* passed to the compressor.
*/
MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_matchState_t *ms)
MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_MatchState_t *ms)
{
return ZSTD_window_hasExtDict(ms->window) ?
ZSTD_extDict :
@@ -1053,7 +1154,9 @@ MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window,
* The least significant cycleLog bits of the indices must remain the same,
* which may be 0. Every index up to maxDist in the past must be valid.
*/
MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog,
U32 maxDist, void const* src)
{
/* preemptive overflow correction:
@@ -1154,7 +1257,7 @@ ZSTD_window_enforceMaxDist(ZSTD_window_t* window,
const void* blockEnd,
U32 maxDist,
U32* loadedDictEndPtr,
const ZSTD_matchState_t** dictMatchStatePtr)
const ZSTD_MatchState_t** dictMatchStatePtr)
{
U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;
@@ -1199,7 +1302,7 @@ ZSTD_checkDictValidity(const ZSTD_window_t* window,
const void* blockEnd,
U32 maxDist,
U32* loadedDictEndPtr,
const ZSTD_matchState_t** dictMatchStatePtr)
const ZSTD_MatchState_t** dictMatchStatePtr)
{
assert(loadedDictEndPtr != NULL);
assert(dictMatchStatePtr != NULL);
@@ -1246,9 +1349,11 @@ MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) {
* forget about the extDict. Handles overlap of the prefix and extDict.
* Returns non-zero if the segment is contiguous.
*/
MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,
void const* src, size_t srcSize,
int forceNonContiguous)
MEM_STATIC
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_window_update(ZSTD_window_t* window,
const void* src, size_t srcSize,
int forceNonContiguous)
{
BYTE const* const ip = (BYTE const*)src;
U32 contiguous = 1;
@@ -1275,8 +1380,9 @@ MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,
/* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */
if ( (ip+srcSize > window->dictBase + window->lowLimit)
& (ip < window->dictBase + window->dictLimit)) {
ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase;
U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx;
size_t const highInputIdx = (size_t)((ip + srcSize) - window->dictBase);
U32 const lowLimitMax = (highInputIdx > (size_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx;
assert(highInputIdx < UINT_MAX);
window->lowLimit = lowLimitMax;
DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit);
}
@@ -1286,7 +1392,7 @@ MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window,
/**
* Returns the lowest allowed match index. It may either be in the ext-dict or the prefix.
*/
MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog)
MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_MatchState_t* ms, U32 curr, unsigned windowLog)
{
U32 const maxDistance = 1U << windowLog;
U32 const lowestValid = ms->window.lowLimit;
@@ -1303,7 +1409,7 @@ MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 curr, u
/**
* Returns the lowest allowed match index in the prefix.
*/
MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_matchState_t* ms, U32 curr, unsigned windowLog)
MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_MatchState_t* ms, U32 curr, unsigned windowLog)
{
U32 const maxDistance = 1U << windowLog;
U32 const lowestValid = ms->window.dictLimit;
@@ -1316,6 +1422,13 @@ MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_matchState_t* ms, U32 curr,
return matchLowest;
}
/* index_safety_check:
* intentional underflow : ensure repIndex isn't overlapping dict + prefix
* @return 1 if values are not overlapping,
* 0 otherwise */
MEM_STATIC int ZSTD_index_overlap_check(const U32 prefixLowestIndex, const U32 repIndex) {
return ((U32)((prefixLowestIndex-1) - repIndex) >= 3);
}
/* debug functions */
@@ -1386,10 +1499,6 @@ MEM_STATIC int ZSTD_comparePackedTags(size_t packedTag1, size_t packedTag2) {
return tag1 == tag2;
}
#if defined (__cplusplus)
}
#endif
/* ===============================================================
* Shared internal declarations
* These prototypes may be called from sources not in lib/compress
@@ -1405,6 +1514,25 @@ size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace,
void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs);
typedef struct {
U32 idx; /* Index in array of ZSTD_Sequence */
U32 posInSequence; /* Position within sequence at idx */
size_t posInSrc; /* Number of bytes given by sequences provided so far */
} ZSTD_SequencePosition;
/* for benchmark */
size_t ZSTD_convertBlockSequences(ZSTD_CCtx* cctx,
const ZSTD_Sequence* const inSeqs, size_t nbSequences,
int const repcodeResolution);
typedef struct {
size_t nbSequences;
size_t blockSize;
size_t litSize;
} BlockSummary;
BlockSummary ZSTD_get1BlockSummary(const ZSTD_Sequence* seqs, size_t nbSeqs);
/* ==============================================================
* Private declarations
* These prototypes shall only be called from within lib/compress
@@ -1416,7 +1544,7 @@ void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs);
* Note: srcSizeHint == 0 means 0!
*/
ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams(
const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize, ZSTD_cParamMode_e mode);
const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize, ZSTD_CParamMode_e mode);
/*! ZSTD_initCStream_internal() :
* Private use only. Init streaming operation.
@@ -1428,7 +1556,7 @@ size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
const ZSTD_CDict* cdict,
const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize);
void ZSTD_resetSeqStore(seqStore_t* ssPtr);
void ZSTD_resetSeqStore(SeqStore_t* ssPtr);
/*! ZSTD_getCParamsFromCDict() :
* as the name implies */
@@ -1467,11 +1595,10 @@ size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);
* This cannot be used when long range matching is enabled.
* Zstd will use these sequences, and pass the literals to a secondary block
* compressor.
* @return : An error code on failure.
* NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory
* access and data corruption.
*/
size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);
void ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq);
/** ZSTD_cycleLog() :
* condition for correct operation : hashLog > 1 */
@@ -1482,33 +1609,10 @@ U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat);
*/
void ZSTD_CCtx_trace(ZSTD_CCtx* cctx, size_t extraCSize);
/* Returns 0 on success, and a ZSTD_error otherwise. This function scans through an array of
* ZSTD_Sequence, storing the sequences it finds, until it reaches a block delimiter.
* Note that the block delimiter must include the last literals of the block.
*/
size_t
ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx,
ZSTD_sequencePosition* seqPos,
const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
const void* src, size_t blockSize, ZSTD_paramSwitch_e externalRepSearch);
/* Returns the number of bytes to move the current read position back by.
* Only non-zero if we ended up splitting a sequence.
* Otherwise, it may return a ZSTD error if something went wrong.
*
* This function will attempt to scan through blockSize bytes
* represented by the sequences in @inSeqs,
* storing any (partial) sequences.
*
* Occasionally, we may want to change the actual number of bytes we consumed from inSeqs to
* avoid splitting a match, or to avoid splitting a match such that it would produce a match
* smaller than MINMATCH. In this case, we return the number of bytes that we didn't read from this block.
*/
size_t
ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
const void* src, size_t blockSize, ZSTD_paramSwitch_e externalRepSearch);
/* Returns 1 if an external sequence producer is registered, otherwise returns 0. */
MEM_STATIC int ZSTD_hasExtSeqProd(const ZSTD_CCtx_params* params) {
return params->extSeqProdFunc != NULL;
}
/* ===============================================================
* Deprecated definitions that are still used internally to avoid
Utilities/cmzstd/lib/compress/zstd_compress_literals.c
+1 -1
View File
@@ -140,7 +140,7 @@ size_t ZSTD_compressLiterals (
size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
BYTE* const ostart = (BYTE*)dst;
U32 singleStream = srcSize < 256;
symbolEncodingType_e hType = set_compressed;
SymbolEncodingType_e hType = set_compressed;
size_t cLitSize;
DEBUGLOG(5,"ZSTD_compressLiterals (disableLiteralCompression=%i, srcSize=%u, dstCapacity=%zu)",
Utilities/cmzstd/lib/compress/zstd_compress_sequences.c
+7 -7
View File
@@ -153,13 +153,13 @@ size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,
return cost >> 8;
}
symbolEncodingType_e
SymbolEncodingType_e
ZSTD_selectEncodingType(
FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
FSE_CTable const* prevCTable,
short const* defaultNorm, U32 defaultNormLog,
ZSTD_defaultPolicy_e const isDefaultAllowed,
ZSTD_DefaultPolicy_e const isDefaultAllowed,
ZSTD_strategy const strategy)
{
ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);
@@ -241,7 +241,7 @@ typedef struct {
size_t
ZSTD_buildCTable(void* dst, size_t dstCapacity,
FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
FSE_CTable* nextCTable, U32 FSELog, SymbolEncodingType_e type,
unsigned* count, U32 max,
const BYTE* codeTable, size_t nbSeq,
const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
@@ -293,7 +293,7 @@ ZSTD_encodeSequences_body(
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
seqDef const* sequences, size_t nbSeq, int longOffsets)
SeqDef const* sequences, size_t nbSeq, int longOffsets)
{
BIT_CStream_t blockStream;
FSE_CState_t stateMatchLength;
@@ -387,7 +387,7 @@ ZSTD_encodeSequences_default(
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
seqDef const* sequences, size_t nbSeq, int longOffsets)
SeqDef const* sequences, size_t nbSeq, int longOffsets)
{
return ZSTD_encodeSequences_body(dst, dstCapacity,
CTable_MatchLength, mlCodeTable,
@@ -405,7 +405,7 @@ ZSTD_encodeSequences_bmi2(
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
seqDef const* sequences, size_t nbSeq, int longOffsets)
SeqDef const* sequences, size_t nbSeq, int longOffsets)
{
return ZSTD_encodeSequences_body(dst, dstCapacity,
CTable_MatchLength, mlCodeTable,
@@ -421,7 +421,7 @@ size_t ZSTD_encodeSequences(
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)
SeqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)
{
DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity);
#if DYNAMIC_BMI2
Utilities/cmzstd/lib/compress/zstd_compress_sequences.h
+7 -6
View File
@@ -11,26 +11,27 @@
#ifndef ZSTD_COMPRESS_SEQUENCES_H
#define ZSTD_COMPRESS_SEQUENCES_H
#include "zstd_compress_internal.h" /* SeqDef */
#include "../common/fse.h" /* FSE_repeat, FSE_CTable */
#include "../common/zstd_internal.h" /* symbolEncodingType_e, ZSTD_strategy */
#include "../common/zstd_internal.h" /* SymbolEncodingType_e, ZSTD_strategy */
typedef enum {
ZSTD_defaultDisallowed = 0,
ZSTD_defaultAllowed = 1
} ZSTD_defaultPolicy_e;
} ZSTD_DefaultPolicy_e;
symbolEncodingType_e
SymbolEncodingType_e
ZSTD_selectEncodingType(
FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
FSE_CTable const* prevCTable,
short const* defaultNorm, U32 defaultNormLog,
ZSTD_defaultPolicy_e const isDefaultAllowed,
ZSTD_DefaultPolicy_e const isDefaultAllowed,
ZSTD_strategy const strategy);
size_t
ZSTD_buildCTable(void* dst, size_t dstCapacity,
FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
FSE_CTable* nextCTable, U32 FSELog, SymbolEncodingType_e type,
unsigned* count, U32 max,
const BYTE* codeTable, size_t nbSeq,
const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
@@ -42,7 +43,7 @@ size_t ZSTD_encodeSequences(
FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2);
SeqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2);
size_t ZSTD_fseBitCost(
FSE_CTable const* ctable,
Utilities/cmzstd/lib/compress/zstd_compress_superblock.c
+233 -122
View File
@@ -51,7 +51,7 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
BYTE* const oend = ostart + dstSize;
BYTE* op = ostart + lhSize;
U32 const singleStream = lhSize == 3;
symbolEncodingType_e hType = writeEntropy ? hufMetadata->hType : set_repeat;
SymbolEncodingType_e hType = writeEntropy ? hufMetadata->hType : set_repeat;
size_t cLitSize = 0;
DEBUGLOG(5, "ZSTD_compressSubBlock_literal (litSize=%zu, lhSize=%zu, writeEntropy=%d)", litSize, lhSize, writeEntropy);
@@ -76,8 +76,8 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
}
{ int const flags = bmi2 ? HUF_flags_bmi2 : 0;
const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, oend-op, literals, litSize, hufTable, flags)
: HUF_compress4X_usingCTable(op, oend-op, literals, litSize, hufTable, flags);
const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags)
: HUF_compress4X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags);
op += cSize;
cLitSize += cSize;
if (cSize == 0 || ERR_isError(cSize)) {
@@ -102,7 +102,7 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
switch(lhSize)
{
case 3: /* 2 - 2 - 10 - 10 */
{ U32 const lhc = hType + ((!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14);
{ U32 const lhc = hType + ((U32)(!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14);
MEM_writeLE24(ostart, lhc);
break;
}
@@ -122,30 +122,30 @@ ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable,
}
*entropyWritten = 1;
DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart));
return op-ostart;
return (size_t)(op-ostart);
}
static size_t
ZSTD_seqDecompressedSize(seqStore_t const* seqStore,
const seqDef* sequences, size_t nbSeq,
size_t litSize, int lastSequence)
ZSTD_seqDecompressedSize(SeqStore_t const* seqStore,
const SeqDef* sequences, size_t nbSeqs,
size_t litSize, int lastSubBlock)
{
const seqDef* const sstart = sequences;
const seqDef* const send = sequences + nbSeq;
const seqDef* sp = sstart;
size_t matchLengthSum = 0;
size_t litLengthSum = 0;
(void)(litLengthSum); /* suppress unused variable warning on some environments */
while (send-sp > 0) {
ZSTD_sequenceLength const seqLen = ZSTD_getSequenceLength(seqStore, sp);
size_t n;
for (n=0; n<nbSeqs; n++) {
const ZSTD_SequenceLength seqLen = ZSTD_getSequenceLength(seqStore, sequences+n);
litLengthSum += seqLen.litLength;
matchLengthSum += seqLen.matchLength;
sp++;
}
assert(litLengthSum <= litSize);
if (!lastSequence) {
DEBUGLOG(5, "ZSTD_seqDecompressedSize: %u sequences from %p: %u literals + %u matchlength",
(unsigned)nbSeqs, (const void*)sequences,
(unsigned)litLengthSum, (unsigned)matchLengthSum);
if (!lastSubBlock)
assert(litLengthSum == litSize);
}
else
assert(litLengthSum <= litSize);
(void)litLengthSum;
return matchLengthSum + litSize;
}
@@ -162,7 +162,7 @@ ZSTD_seqDecompressedSize(seqStore_t const* seqStore,
static size_t
ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
const ZSTD_fseCTablesMetadata_t* fseMetadata,
const seqDef* sequences, size_t nbSeq,
const SeqDef* sequences, size_t nbSeq,
const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode,
const ZSTD_CCtx_params* cctxParams,
void* dst, size_t dstCapacity,
@@ -180,14 +180,14 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
/* Sequences Header */
RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
dstSize_tooSmall, "");
if (nbSeq < 0x7F)
if (nbSeq < 128)
*op++ = (BYTE)nbSeq;
else if (nbSeq < LONGNBSEQ)
op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;
else
op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;
if (nbSeq==0) {
return op - ostart;
return (size_t)(op - ostart);
}
/* seqHead : flags for FSE encoding type */
@@ -209,7 +209,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
}
{ size_t const bitstreamSize = ZSTD_encodeSequences(
op, oend - op,
op, (size_t)(oend - op),
fseTables->matchlengthCTable, mlCode,
fseTables->offcodeCTable, ofCode,
fseTables->litlengthCTable, llCode,
@@ -253,7 +253,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
#endif
*entropyWritten = 1;
return op - ostart;
return (size_t)(op - ostart);
}
/** ZSTD_compressSubBlock() :
@@ -262,7 +262,7 @@ ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables,
* Or 0 if it failed to compress. */
static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
const seqDef* sequences, size_t nbSeq,
const SeqDef* sequences, size_t nbSeq,
const BYTE* literals, size_t litSize,
const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode,
const ZSTD_CCtx_params* cctxParams,
@@ -279,7 +279,8 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
litSize, nbSeq, writeLitEntropy, writeSeqEntropy, lastBlock);
{ size_t cLitSize = ZSTD_compressSubBlock_literal((const HUF_CElt*)entropy->huf.CTable,
&entropyMetadata->hufMetadata, literals, litSize,
op, oend-op, bmi2, writeLitEntropy, litEntropyWritten);
op, (size_t)(oend-op),
bmi2, writeLitEntropy, litEntropyWritten);
FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed");
if (cLitSize == 0) return 0;
op += cLitSize;
@@ -289,18 +290,18 @@ static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy,
sequences, nbSeq,
llCode, mlCode, ofCode,
cctxParams,
op, oend-op,
op, (size_t)(oend-op),
bmi2, writeSeqEntropy, seqEntropyWritten);
FORWARD_IF_ERROR(cSeqSize, "ZSTD_compressSubBlock_sequences failed");
if (cSeqSize == 0) return 0;
op += cSeqSize;
}
/* Write block header */
{ size_t cSize = (op-ostart)-ZSTD_blockHeaderSize;
{ size_t cSize = (size_t)(op-ostart) - ZSTD_blockHeaderSize;
U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
MEM_writeLE24(ostart, cBlockHeader24);
}
return op-ostart;
return (size_t)(op-ostart);
}
static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize,
@@ -326,7 +327,7 @@ static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t lit
return 0;
}
static size_t ZSTD_estimateSubBlockSize_symbolType(symbolEncodingType_e type,
static size_t ZSTD_estimateSubBlockSize_symbolType(SymbolEncodingType_e type,
const BYTE* codeTable, unsigned maxCode,
size_t nbSeq, const FSE_CTable* fseCTable,
const U8* additionalBits,
@@ -389,7 +390,11 @@ static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable,
return cSeqSizeEstimate + sequencesSectionHeaderSize;
}
static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
typedef struct {
size_t estLitSize;
size_t estBlockSize;
} EstimatedBlockSize;
static EstimatedBlockSize ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
const BYTE* ofCodeTable,
const BYTE* llCodeTable,
const BYTE* mlCodeTable,
@@ -397,15 +402,17 @@ static size_t ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize,
const ZSTD_entropyCTables_t* entropy,
const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
void* workspace, size_t wkspSize,
int writeLitEntropy, int writeSeqEntropy) {
size_t cSizeEstimate = 0;
cSizeEstimate += ZSTD_estimateSubBlockSize_literal(literals, litSize,
&entropy->huf, &entropyMetadata->hufMetadata,
workspace, wkspSize, writeLitEntropy);
cSizeEstimate += ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
int writeLitEntropy, int writeSeqEntropy)
{
EstimatedBlockSize ebs;
ebs.estLitSize = ZSTD_estimateSubBlockSize_literal(literals, litSize,
&entropy->huf, &entropyMetadata->hufMetadata,
workspace, wkspSize, writeLitEntropy);
ebs.estBlockSize = ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable,
nbSeq, &entropy->fse, &entropyMetadata->fseMetadata,
workspace, wkspSize, writeSeqEntropy);
return cSizeEstimate + ZSTD_blockHeaderSize;
ebs.estBlockSize += ebs.estLitSize + ZSTD_blockHeaderSize;
return ebs;
}
static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata)
@@ -419,14 +426,57 @@ static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMe
return 0;
}
static size_t countLiterals(SeqStore_t const* seqStore, const SeqDef* sp, size_t seqCount)
{
size_t n, total = 0;
assert(sp != NULL);
for (n=0; n<seqCount; n++) {
total += ZSTD_getSequenceLength(seqStore, sp+n).litLength;
}
DEBUGLOG(6, "countLiterals for %zu sequences from %p => %zu bytes", seqCount, (const void*)sp, total);
return total;
}
#define BYTESCALE 256
static size_t sizeBlockSequences(const SeqDef* sp, size_t nbSeqs,
size_t targetBudget, size_t avgLitCost, size_t avgSeqCost,
int firstSubBlock)
{
size_t n, budget = 0, inSize=0;
/* entropy headers */
size_t const headerSize = (size_t)firstSubBlock * 120 * BYTESCALE; /* generous estimate */
assert(firstSubBlock==0 || firstSubBlock==1);
budget += headerSize;
/* first sequence => at least one sequence*/
budget += sp[0].litLength * avgLitCost + avgSeqCost;
if (budget > targetBudget) return 1;
inSize = sp[0].litLength + (sp[0].mlBase+MINMATCH);
/* loop over sequences */
for (n=1; n<nbSeqs; n++) {
size_t currentCost = sp[n].litLength * avgLitCost + avgSeqCost;
budget += currentCost;
inSize += sp[n].litLength + (sp[n].mlBase+MINMATCH);
/* stop when sub-block budget is reached */
if ( (budget > targetBudget)
/* though continue to expand until the sub-block is deemed compressible */
&& (budget < inSize * BYTESCALE) )
break;
}
return n;
}
/** ZSTD_compressSubBlock_multi() :
* Breaks super-block into multiple sub-blocks and compresses them.
* Entropy will be written to the first block.
* The following blocks will use repeat mode to compress.
* All sub-blocks are compressed blocks (no raw or rle blocks).
* @return : compressed size of the super block (which is multiple ZSTD blocks)
* Or 0 if it failed to compress. */
static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
* Entropy will be written into the first block.
* The following blocks use repeat_mode to compress.
* Sub-blocks are all compressed, except the last one when beneficial.
* @return : compressed size of the super block (which features multiple ZSTD blocks)
* or 0 if it failed to compress. */
static size_t ZSTD_compressSubBlock_multi(const SeqStore_t* seqStorePtr,
const ZSTD_compressedBlockState_t* prevCBlock,
ZSTD_compressedBlockState_t* nextCBlock,
const ZSTD_entropyCTablesMetadata_t* entropyMetadata,
@@ -436,12 +486,14 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
const int bmi2, U32 lastBlock,
void* workspace, size_t wkspSize)
{
const seqDef* const sstart = seqStorePtr->sequencesStart;
const seqDef* const send = seqStorePtr->sequences;
const seqDef* sp = sstart;
const SeqDef* const sstart = seqStorePtr->sequencesStart;
const SeqDef* const send = seqStorePtr->sequences;
const SeqDef* sp = sstart; /* tracks progresses within seqStorePtr->sequences */
size_t const nbSeqs = (size_t)(send - sstart);
const BYTE* const lstart = seqStorePtr->litStart;
const BYTE* const lend = seqStorePtr->lit;
const BYTE* lp = lstart;
size_t const nbLiterals = (size_t)(lend - lstart);
BYTE const* ip = (BYTE const*)src;
BYTE const* const iend = ip + srcSize;
BYTE* const ostart = (BYTE*)dst;
@@ -450,97 +502,153 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
const BYTE* llCodePtr = seqStorePtr->llCode;
const BYTE* mlCodePtr = seqStorePtr->mlCode;
const BYTE* ofCodePtr = seqStorePtr->ofCode;
size_t targetCBlockSize = cctxParams->targetCBlockSize;
size_t litSize, seqCount;
int writeLitEntropy = entropyMetadata->hufMetadata.hType == set_compressed;
size_t const minTarget = ZSTD_TARGETCBLOCKSIZE_MIN; /* enforce minimum size, to reduce undesirable side effects */
size_t const targetCBlockSize = MAX(minTarget, cctxParams->targetCBlockSize);
int writeLitEntropy = (entropyMetadata->hufMetadata.hType == set_compressed);
int writeSeqEntropy = 1;
int lastSequence = 0;
DEBUGLOG(5, "ZSTD_compressSubBlock_multi (litSize=%u, nbSeq=%u)",
(unsigned)(lend-lp), (unsigned)(send-sstart));
DEBUGLOG(5, "ZSTD_compressSubBlock_multi (srcSize=%u, litSize=%u, nbSeq=%u)",
(unsigned)srcSize, (unsigned)(lend-lstart), (unsigned)(send-sstart));
litSize = 0;
seqCount = 0;
do {
size_t cBlockSizeEstimate = 0;
if (sstart == send) {
lastSequence = 1;
} else {
const seqDef* const sequence = sp + seqCount;
lastSequence = sequence == send - 1;
litSize += ZSTD_getSequenceLength(seqStorePtr, sequence).litLength;
seqCount++;
/* let's start by a general estimation for the full block */
if (nbSeqs > 0) {
EstimatedBlockSize const ebs =
ZSTD_estimateSubBlockSize(lp, nbLiterals,
ofCodePtr, llCodePtr, mlCodePtr, nbSeqs,
&nextCBlock->entropy, entropyMetadata,
workspace, wkspSize,
writeLitEntropy, writeSeqEntropy);
/* quick estimation */
size_t const avgLitCost = nbLiterals ? (ebs.estLitSize * BYTESCALE) / nbLiterals : BYTESCALE;
size_t const avgSeqCost = ((ebs.estBlockSize - ebs.estLitSize) * BYTESCALE) / nbSeqs;
const size_t nbSubBlocks = MAX((ebs.estBlockSize + (targetCBlockSize/2)) / targetCBlockSize, 1);
size_t n, avgBlockBudget, blockBudgetSupp=0;
avgBlockBudget = (ebs.estBlockSize * BYTESCALE) / nbSubBlocks;
DEBUGLOG(5, "estimated fullblock size=%u bytes ; avgLitCost=%.2f ; avgSeqCost=%.2f ; targetCBlockSize=%u, nbSubBlocks=%u ; avgBlockBudget=%.0f bytes",
(unsigned)ebs.estBlockSize, (double)avgLitCost/BYTESCALE, (double)avgSeqCost/BYTESCALE,
(unsigned)targetCBlockSize, (unsigned)nbSubBlocks, (double)avgBlockBudget/BYTESCALE);
/* simplification: if estimates states that the full superblock doesn't compress, just bail out immediately
* this will result in the production of a single uncompressed block covering @srcSize.*/
if (ebs.estBlockSize > srcSize) return 0;
/* compress and write sub-blocks */
assert(nbSubBlocks>0);
for (n=0; n < nbSubBlocks-1; n++) {
/* determine nb of sequences for current sub-block + nbLiterals from next sequence */
size_t const seqCount = sizeBlockSequences(sp, (size_t)(send-sp),
avgBlockBudget + blockBudgetSupp, avgLitCost, avgSeqCost, n==0);
/* if reached last sequence : break to last sub-block (simplification) */
assert(seqCount <= (size_t)(send-sp));
if (sp + seqCount == send) break;
assert(seqCount > 0);
/* compress sub-block */
{ int litEntropyWritten = 0;
int seqEntropyWritten = 0;
size_t litSize = countLiterals(seqStorePtr, sp, seqCount);
const size_t decompressedSize =
ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 0);
size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
sp, seqCount,
lp, litSize,
llCodePtr, mlCodePtr, ofCodePtr,
cctxParams,
op, (size_t)(oend-op),
bmi2, writeLitEntropy, writeSeqEntropy,
&litEntropyWritten, &seqEntropyWritten,
0);
FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
/* check compressibility, update state components */
if (cSize > 0 && cSize < decompressedSize) {
DEBUGLOG(5, "Committed sub-block compressing %u bytes => %u bytes",
(unsigned)decompressedSize, (unsigned)cSize);
assert(ip + decompressedSize <= iend);
ip += decompressedSize;
lp += litSize;
op += cSize;
llCodePtr += seqCount;
mlCodePtr += seqCount;
ofCodePtr += seqCount;
/* Entropy only needs to be written once */
if (litEntropyWritten) {
writeLitEntropy = 0;
}
if (seqEntropyWritten) {
writeSeqEntropy = 0;
}
sp += seqCount;
blockBudgetSupp = 0;
} }
/* otherwise : do not compress yet, coalesce current sub-block with following one */
}
if (lastSequence) {
assert(lp <= lend);
assert(litSize <= (size_t)(lend - lp));
litSize = (size_t)(lend - lp);
}
/* I think there is an optimization opportunity here.
* Calling ZSTD_estimateSubBlockSize for every sequence can be wasteful
* since it recalculates estimate from scratch.
* For example, it would recount literal distribution and symbol codes every time.
*/
cBlockSizeEstimate = ZSTD_estimateSubBlockSize(lp, litSize, ofCodePtr, llCodePtr, mlCodePtr, seqCount,
&nextCBlock->entropy, entropyMetadata,
workspace, wkspSize, writeLitEntropy, writeSeqEntropy);
if (cBlockSizeEstimate > targetCBlockSize || lastSequence) {
int litEntropyWritten = 0;
int seqEntropyWritten = 0;
const size_t decompressedSize = ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, lastSequence);
const size_t cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
sp, seqCount,
lp, litSize,
llCodePtr, mlCodePtr, ofCodePtr,
cctxParams,
op, oend-op,
bmi2, writeLitEntropy, writeSeqEntropy,
&litEntropyWritten, &seqEntropyWritten,
lastBlock && lastSequence);
FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
if (cSize > 0 && cSize < decompressedSize) {
DEBUGLOG(5, "Committed the sub-block");
assert(ip + decompressedSize <= iend);
ip += decompressedSize;
sp += seqCount;
lp += litSize;
op += cSize;
llCodePtr += seqCount;
mlCodePtr += seqCount;
ofCodePtr += seqCount;
litSize = 0;
seqCount = 0;
/* Entropy only needs to be written once */
if (litEntropyWritten) {
writeLitEntropy = 0;
}
if (seqEntropyWritten) {
writeSeqEntropy = 0;
}
} /* if (nbSeqs > 0) */
/* write last block */
DEBUGLOG(5, "Generate last sub-block: %u sequences remaining", (unsigned)(send - sp));
{ int litEntropyWritten = 0;
int seqEntropyWritten = 0;
size_t litSize = (size_t)(lend - lp);
size_t seqCount = (size_t)(send - sp);
const size_t decompressedSize =
ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 1);
size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata,
sp, seqCount,
lp, litSize,
llCodePtr, mlCodePtr, ofCodePtr,
cctxParams,
op, (size_t)(oend-op),
bmi2, writeLitEntropy, writeSeqEntropy,
&litEntropyWritten, &seqEntropyWritten,
lastBlock);
FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed");
/* update pointers, the nb of literals borrowed from next sequence must be preserved */
if (cSize > 0 && cSize < decompressedSize) {
DEBUGLOG(5, "Last sub-block compressed %u bytes => %u bytes",
(unsigned)decompressedSize, (unsigned)cSize);
assert(ip + decompressedSize <= iend);
ip += decompressedSize;
lp += litSize;
op += cSize;
llCodePtr += seqCount;
mlCodePtr += seqCount;
ofCodePtr += seqCount;
/* Entropy only needs to be written once */
if (litEntropyWritten) {
writeLitEntropy = 0;
}
if (seqEntropyWritten) {
writeSeqEntropy = 0;
}
sp += seqCount;
}
} while (!lastSequence);
}
if (writeLitEntropy) {
DEBUGLOG(5, "ZSTD_compressSubBlock_multi has literal entropy tables unwritten");
DEBUGLOG(5, "Literal entropy tables were never written");
ZSTD_memcpy(&nextCBlock->entropy.huf, &prevCBlock->entropy.huf, sizeof(prevCBlock->entropy.huf));
}
if (writeSeqEntropy && ZSTD_needSequenceEntropyTables(&entropyMetadata->fseMetadata)) {
/* If we haven't written our entropy tables, then we've violated our contract and
* must emit an uncompressed block.
*/
DEBUGLOG(5, "ZSTD_compressSubBlock_multi has sequence entropy tables unwritten");
DEBUGLOG(5, "Sequence entropy tables were never written => cancel, emit an uncompressed block");
return 0;
}
if (ip < iend) {
size_t const cSize = ZSTD_noCompressBlock(op, oend - op, ip, iend - ip, lastBlock);
DEBUGLOG(5, "ZSTD_compressSubBlock_multi last sub-block uncompressed, %zu bytes", (size_t)(iend - ip));
/* some data left : last part of the block sent uncompressed */
size_t const rSize = (size_t)((iend - ip));
size_t const cSize = ZSTD_noCompressBlock(op, (size_t)(oend - op), ip, rSize, lastBlock);
DEBUGLOG(5, "Generate last uncompressed sub-block of %u bytes", (unsigned)(rSize));
FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed");
assert(cSize != 0);
op += cSize;
/* We have to regenerate the repcodes because we've skipped some sequences */
if (sp < send) {
seqDef const* seq;
repcodes_t rep;
const SeqDef* seq;
Repcodes_t rep;
ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep));
for (seq = sstart; seq < sp; ++seq) {
ZSTD_updateRep(rep.rep, seq->offBase, ZSTD_getSequenceLength(seqStorePtr, seq).litLength == 0);
@@ -548,14 +656,17 @@ static size_t ZSTD_compressSubBlock_multi(const seqStore_t* seqStorePtr,
ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep));
}
}
DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed");
return op-ostart;
DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed all subBlocks: total compressed size = %u",
(unsigned)(op-ostart));
return (size_t)(op-ostart);
}
size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
void* dst, size_t dstCapacity,
void const* src, size_t srcSize,
unsigned lastBlock) {
const void* src, size_t srcSize,
unsigned lastBlock)
{
ZSTD_entropyCTablesMetadata_t entropyMetadata;
FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(&zc->seqStore,
@@ -563,7 +674,7 @@ size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
&zc->blockState.nextCBlock->entropy,
&zc->appliedParams,
&entropyMetadata,
zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */), "");
zc->tmpWorkspace, zc->tmpWkspSize /* statically allocated in resetCCtx */), "");
return ZSTD_compressSubBlock_multi(&zc->seqStore,
zc->blockState.prevCBlock,
@@ -573,5 +684,5 @@ size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc,
dst, dstCapacity,
src, srcSize,
zc->bmi2, lastBlock,
zc->entropyWorkspace, ENTROPY_WORKSPACE_SIZE /* statically allocated in resetCCtx */);
zc->tmpWorkspace, zc->tmpWkspSize /* statically allocated in resetCCtx */);
}
Utilities/cmzstd/lib/compress/zstd_cwksp.h
+60 -37
View File
@@ -17,10 +17,7 @@
#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customFree */
#include "../common/zstd_internal.h"
#include "../common/portability_macros.h"
#if defined (__cplusplus)
extern "C" {
#endif
#include "../common/compiler.h" /* ZS2_isPower2 */
/*-*************************************
* Constants
@@ -192,6 +189,7 @@ MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
{
intptr_t const offset = __msan_test_shadow(ws->initOnceStart,
(U8*)ZSTD_cwksp_initialAllocStart(ws) - (U8*)ws->initOnceStart);
(void)offset;
#if defined(ZSTD_MSAN_PRINT)
if(offset!=-1) {
__msan_print_shadow((U8*)ws->initOnceStart + offset - 8, 32);
@@ -205,9 +203,9 @@ MEM_STATIC void ZSTD_cwksp_assert_internal_consistency(ZSTD_cwksp* ws) {
/**
* Align must be a power of 2.
*/
MEM_STATIC size_t ZSTD_cwksp_align(size_t size, size_t const align) {
MEM_STATIC size_t ZSTD_cwksp_align(size_t size, size_t align) {
size_t const mask = align - 1;
assert((align & mask) == 0);
assert(ZSTD_isPower2(align));
return (size + mask) & ~mask;
}
@@ -221,7 +219,7 @@ MEM_STATIC size_t ZSTD_cwksp_align(size_t size, size_t const align) {
* to figure out how much space you need for the matchState tables. Everything
* else is though.
*
* Do not use for sizing aligned buffers. Instead, use ZSTD_cwksp_aligned_alloc_size().
* Do not use for sizing aligned buffers. Instead, use ZSTD_cwksp_aligned64_alloc_size().
*/
MEM_STATIC size_t ZSTD_cwksp_alloc_size(size_t size) {
if (size == 0)
@@ -233,12 +231,16 @@ MEM_STATIC size_t ZSTD_cwksp_alloc_size(size_t size) {
#endif
}
MEM_STATIC size_t ZSTD_cwksp_aligned_alloc_size(size_t size, size_t alignment) {
return ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(size, alignment));
}
/**
* Returns an adjusted alloc size that is the nearest larger multiple of 64 bytes.
* Used to determine the number of bytes required for a given "aligned".
*/
MEM_STATIC size_t ZSTD_cwksp_aligned_alloc_size(size_t size) {
return ZSTD_cwksp_alloc_size(ZSTD_cwksp_align(size, ZSTD_CWKSP_ALIGNMENT_BYTES));
MEM_STATIC size_t ZSTD_cwksp_aligned64_alloc_size(size_t size) {
return ZSTD_cwksp_aligned_alloc_size(size, ZSTD_CWKSP_ALIGNMENT_BYTES);
}
/**
@@ -261,7 +263,7 @@ MEM_STATIC size_t ZSTD_cwksp_slack_space_required(void) {
MEM_STATIC size_t ZSTD_cwksp_bytes_to_align_ptr(void* ptr, const size_t alignBytes) {
size_t const alignBytesMask = alignBytes - 1;
size_t const bytes = (alignBytes - ((size_t)ptr & (alignBytesMask))) & alignBytesMask;
assert((alignBytes & alignBytesMask) == 0);
assert(ZSTD_isPower2(alignBytes));
assert(bytes < alignBytes);
return bytes;
}
@@ -270,8 +272,12 @@ MEM_STATIC size_t ZSTD_cwksp_bytes_to_align_ptr(void* ptr, const size_t alignByt
* Returns the initial value for allocStart which is used to determine the position from
* which we can allocate from the end of the workspace.
*/
MEM_STATIC void* ZSTD_cwksp_initialAllocStart(ZSTD_cwksp* ws) {
return (void*)((size_t)ws->workspaceEnd & ~(ZSTD_CWKSP_ALIGNMENT_BYTES-1));
MEM_STATIC void* ZSTD_cwksp_initialAllocStart(ZSTD_cwksp* ws)
{
char* endPtr = (char*)ws->workspaceEnd;
assert(ZSTD_isPower2(ZSTD_CWKSP_ALIGNMENT_BYTES));
endPtr = endPtr - ((size_t)endPtr % ZSTD_CWKSP_ALIGNMENT_BYTES);
return (void*)endPtr;
}
/**
@@ -286,7 +292,7 @@ ZSTD_cwksp_reserve_internal_buffer_space(ZSTD_cwksp* ws, size_t const bytes)
{
void* const alloc = (BYTE*)ws->allocStart - bytes;
void* const bottom = ws->tableEnd;
DEBUGLOG(5, "cwksp: reserving %p %zd bytes, %zd bytes remaining",
DEBUGLOG(5, "cwksp: reserving [0x%p]:%zd bytes; %zd bytes remaining",
alloc, bytes, ZSTD_cwksp_available_space(ws) - bytes);
ZSTD_cwksp_assert_internal_consistency(ws);
assert(alloc >= bottom);
@@ -403,7 +409,7 @@ MEM_STATIC void* ZSTD_cwksp_reserve_aligned_init_once(ZSTD_cwksp* ws, size_t byt
{
size_t const alignedBytes = ZSTD_cwksp_align(bytes, ZSTD_CWKSP_ALIGNMENT_BYTES);
void* ptr = ZSTD_cwksp_reserve_internal(ws, alignedBytes, ZSTD_cwksp_alloc_aligned_init_once);
assert(((size_t)ptr & (ZSTD_CWKSP_ALIGNMENT_BYTES-1))== 0);
assert(((size_t)ptr & (ZSTD_CWKSP_ALIGNMENT_BYTES-1)) == 0);
if(ptr && ptr < ws->initOnceStart) {
/* We assume the memory following the current allocation is either:
* 1. Not usable as initOnce memory (end of workspace)
@@ -423,17 +429,18 @@ MEM_STATIC void* ZSTD_cwksp_reserve_aligned_init_once(ZSTD_cwksp* ws, size_t byt
/**
* Reserves and returns memory sized on and aligned on ZSTD_CWKSP_ALIGNMENT_BYTES (64 bytes).
*/
MEM_STATIC void* ZSTD_cwksp_reserve_aligned(ZSTD_cwksp* ws, size_t bytes)
MEM_STATIC void* ZSTD_cwksp_reserve_aligned64(ZSTD_cwksp* ws, size_t bytes)
{
void* ptr = ZSTD_cwksp_reserve_internal(ws, ZSTD_cwksp_align(bytes, ZSTD_CWKSP_ALIGNMENT_BYTES),
ZSTD_cwksp_alloc_aligned);
assert(((size_t)ptr & (ZSTD_CWKSP_ALIGNMENT_BYTES-1))== 0);
void* const ptr = ZSTD_cwksp_reserve_internal(ws,
ZSTD_cwksp_align(bytes, ZSTD_CWKSP_ALIGNMENT_BYTES),
ZSTD_cwksp_alloc_aligned);
assert(((size_t)ptr & (ZSTD_CWKSP_ALIGNMENT_BYTES-1)) == 0);
return ptr;
}
/**
* Aligned on 64 bytes. These buffers have the special property that
* their values remain constrained, allowing us to re-use them without
* their values remain constrained, allowing us to reuse them without
* memset()-ing them.
*/
MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes)
@@ -473,7 +480,7 @@ MEM_STATIC void* ZSTD_cwksp_reserve_table(ZSTD_cwksp* ws, size_t bytes)
#endif
assert((bytes & (ZSTD_CWKSP_ALIGNMENT_BYTES-1)) == 0);
assert(((size_t)alloc & (ZSTD_CWKSP_ALIGNMENT_BYTES-1))== 0);
assert(((size_t)alloc & (ZSTD_CWKSP_ALIGNMENT_BYTES-1)) == 0);
return alloc;
}
@@ -519,13 +526,27 @@ MEM_STATIC void* ZSTD_cwksp_reserve_object(ZSTD_cwksp* ws, size_t bytes)
return alloc;
}
/**
* with alignment control
* Note : should happen only once, at workspace first initialization
*/
MEM_STATIC void* ZSTD_cwksp_reserve_object_aligned(ZSTD_cwksp* ws, size_t byteSize, size_t alignment)
{
size_t const mask = alignment - 1;
size_t const surplus = (alignment > sizeof(void*)) ? alignment - sizeof(void*) : 0;
void* const start = ZSTD_cwksp_reserve_object(ws, byteSize + surplus);
if (start == NULL) return NULL;
if (surplus == 0) return start;
assert(ZSTD_isPower2(alignment));
return (void*)(((size_t)start + surplus) & ~mask);
}
MEM_STATIC void ZSTD_cwksp_mark_tables_dirty(ZSTD_cwksp* ws)
{
DEBUGLOG(4, "cwksp: ZSTD_cwksp_mark_tables_dirty");
#if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE)
/* To validate that the table re-use logic is sound, and that we don't
/* To validate that the table reuse logic is sound, and that we don't
* access table space that we haven't cleaned, we re-"poison" the table
* space every time we mark it dirty.
* Since tableValidEnd space and initOnce space may overlap we don't poison
@@ -576,7 +597,8 @@ MEM_STATIC void ZSTD_cwksp_clean_tables(ZSTD_cwksp* ws) {
* Invalidates table allocations.
* All other allocations remain valid.
*/
MEM_STATIC void ZSTD_cwksp_clear_tables(ZSTD_cwksp* ws) {
MEM_STATIC void ZSTD_cwksp_clear_tables(ZSTD_cwksp* ws)
{
DEBUGLOG(4, "cwksp: clearing tables!");
#if ZSTD_ADDRESS_SANITIZER && !defined (ZSTD_ASAN_DONT_POISON_WORKSPACE)
@@ -602,9 +624,9 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
DEBUGLOG(4, "cwksp: clearing!");
#if ZSTD_MEMORY_SANITIZER && !defined (ZSTD_MSAN_DONT_POISON_WORKSPACE)
/* To validate that the context re-use logic is sound, and that we don't
/* To validate that the context reuse logic is sound, and that we don't
* access stuff that this compression hasn't initialized, we re-"poison"
* the workspace except for the areas in which we expect memory re-use
* the workspace except for the areas in which we expect memory reuse
* without initialization (objects, valid tables area and init once
* memory). */
{
@@ -635,6 +657,15 @@ MEM_STATIC void ZSTD_cwksp_clear(ZSTD_cwksp* ws) {
ZSTD_cwksp_assert_internal_consistency(ws);
}
MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) {
return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace);
}
MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) {
return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace)
+ (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart);
}
/**
* The provided workspace takes ownership of the buffer [start, start+size).
* Any existing values in the workspace are ignored (the previously managed
@@ -666,6 +697,11 @@ MEM_STATIC size_t ZSTD_cwksp_create(ZSTD_cwksp* ws, size_t size, ZSTD_customMem
MEM_STATIC void ZSTD_cwksp_free(ZSTD_cwksp* ws, ZSTD_customMem customMem) {
void *ptr = ws->workspace;
DEBUGLOG(4, "cwksp: freeing workspace");
#if ZSTD_MEMORY_SANITIZER && !defined(ZSTD_MSAN_DONT_POISON_WORKSPACE)
if (ptr != NULL && customMem.customFree != NULL) {
__msan_unpoison(ptr, ZSTD_cwksp_sizeof(ws));
}
#endif
ZSTD_memset(ws, 0, sizeof(ZSTD_cwksp));
ZSTD_customFree(ptr, customMem);
}
@@ -679,15 +715,6 @@ MEM_STATIC void ZSTD_cwksp_move(ZSTD_cwksp* dst, ZSTD_cwksp* src) {
ZSTD_memset(src, 0, sizeof(ZSTD_cwksp));
}
MEM_STATIC size_t ZSTD_cwksp_sizeof(const ZSTD_cwksp* ws) {
return (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->workspace);
}
MEM_STATIC size_t ZSTD_cwksp_used(const ZSTD_cwksp* ws) {
return (size_t)((BYTE*)ws->tableEnd - (BYTE*)ws->workspace)
+ (size_t)((BYTE*)ws->workspaceEnd - (BYTE*)ws->allocStart);
}
MEM_STATIC int ZSTD_cwksp_reserve_failed(const ZSTD_cwksp* ws) {
return ws->allocFailed;
}
@@ -735,8 +762,4 @@ MEM_STATIC void ZSTD_cwksp_bump_oversized_duration(
}
}
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_CWKSP_H */
Utilities/cmzstd/lib/compress/zstd_double_fast.c
+73 -53
View File
@@ -11,7 +11,11 @@
#include "zstd_compress_internal.h"
#include "zstd_double_fast.h"
static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms,
#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_fillDoubleHashTableForCDict(ZSTD_MatchState_t* ms,
void const* end, ZSTD_dictTableLoadMethod_e dtlm)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -47,7 +51,9 @@ static void ZSTD_fillDoubleHashTableForCDict(ZSTD_matchState_t* ms,
} }
}
static void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_fillDoubleHashTableForCCtx(ZSTD_MatchState_t* ms,
void const* end, ZSTD_dictTableLoadMethod_e dtlm)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -81,7 +87,7 @@ static void ZSTD_fillDoubleHashTableForCCtx(ZSTD_matchState_t* ms,
} }
}
void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
void ZSTD_fillDoubleHashTable(ZSTD_MatchState_t* ms,
const void* const end,
ZSTD_dictTableLoadMethod_e dtlm,
ZSTD_tableFillPurpose_e tfp)
@@ -95,8 +101,9 @@ void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_doubleFast_noDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, U32 const mls /* template */)
{
ZSTD_compressionParameters const* cParams = &ms->cParams;
@@ -135,9 +142,14 @@ size_t ZSTD_compressBlock_doubleFast_noDict_generic(
const BYTE* matchl0; /* the long match for ip */
const BYTE* matchs0; /* the short match for ip */
const BYTE* matchl1; /* the long match for ip1 */
const BYTE* matchs0_safe; /* matchs0 or safe address */
const BYTE* ip = istart; /* the current position */
const BYTE* ip1; /* the next position */
/* Array of ~random data, should have low probability of matching data
* we load from here instead of from tables, if matchl0/matchl1 are
* invalid indices. Used to avoid unpredictable branches. */
const BYTE dummy[] = {0x12,0x34,0x56,0x78,0x9a,0xbc,0xde,0xf0,0xe2,0xb4};
DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_noDict_generic");
@@ -184,24 +196,29 @@ size_t ZSTD_compressBlock_doubleFast_noDict_generic(
hl1 = ZSTD_hashPtr(ip1, hBitsL, 8);
if (idxl0 > prefixLowestIndex) {
/* idxl0 > prefixLowestIndex is a (somewhat) unpredictable branch.
* However expression below complies into conditional move. Since
* match is unlikely and we only *branch* on idxl0 > prefixLowestIndex
* if there is a match, all branches become predictable. */
{ const BYTE* const matchl0_safe = ZSTD_selectAddr(idxl0, prefixLowestIndex, matchl0, &dummy[0]);
/* check prefix long match */
if (MEM_read64(matchl0) == MEM_read64(ip)) {
if (MEM_read64(matchl0_safe) == MEM_read64(ip) && matchl0_safe == matchl0) {
mLength = ZSTD_count(ip+8, matchl0+8, iend) + 8;
offset = (U32)(ip-matchl0);
while (((ip>anchor) & (matchl0>prefixLowest)) && (ip[-1] == matchl0[-1])) { ip--; matchl0--; mLength++; } /* catch up */
goto _match_found;
}
}
} }
idxl1 = hashLong[hl1];
matchl1 = base + idxl1;
if (idxs0 > prefixLowestIndex) {
/* check prefix short match */
if (MEM_read32(matchs0) == MEM_read32(ip)) {
goto _search_next_long;
}
/* Same optimization as matchl0 above */
matchs0_safe = ZSTD_selectAddr(idxs0, prefixLowestIndex, matchs0, &dummy[0]);
/* check prefix short match */
if(MEM_read32(matchs0_safe) == MEM_read32(ip) && matchs0_safe == matchs0) {
goto _search_next_long;
}
if (ip1 >= nextStep) {
@@ -235,21 +252,23 @@ _cleanup:
_search_next_long:
/* check prefix long +1 match */
if (idxl1 > prefixLowestIndex) {
if (MEM_read64(matchl1) == MEM_read64(ip1)) {
/* short match found: let's check for a longer one */
mLength = ZSTD_count(ip+4, matchs0+4, iend) + 4;
offset = (U32)(ip - matchs0);
/* check long match at +1 position */
if ((idxl1 > prefixLowestIndex) && (MEM_read64(matchl1) == MEM_read64(ip1))) {
size_t const l1len = ZSTD_count(ip1+8, matchl1+8, iend) + 8;
if (l1len > mLength) {
/* use the long match instead */
ip = ip1;
mLength = ZSTD_count(ip+8, matchl1+8, iend) + 8;
mLength = l1len;
offset = (U32)(ip-matchl1);
while (((ip>anchor) & (matchl1>prefixLowest)) && (ip[-1] == matchl1[-1])) { ip--; matchl1--; mLength++; } /* catch up */
goto _match_found;
matchs0 = matchl1;
}
}
/* if no long +1 match, explore the short match we found */
mLength = ZSTD_count(ip+4, matchs0+4, iend) + 4;
offset = (U32)(ip - matchs0);
while (((ip>anchor) & (matchs0>prefixLowest)) && (ip[-1] == matchs0[-1])) { ip--; matchs0--; mLength++; } /* catch up */
while (((ip>anchor) & (matchs0>prefixLowest)) && (ip[-1] == matchs0[-1])) { ip--; matchs0--; mLength++; } /* complete backward */
/* fall-through */
@@ -305,8 +324,9 @@ _match_stored:
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize,
U32 const mls /* template */)
{
@@ -327,7 +347,7 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=rep[0], offset_2=rep[1];
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_MatchState_t* const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dictCParams = &dms->cParams;
const U32* const dictHashLong = dms->hashTable;
const U32* const dictHashSmall = dms->chainTable;
@@ -348,8 +368,8 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
if (ms->prefetchCDictTables) {
size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32);
size_t const chainTableBytes = (((size_t)1) << dictCParams->chainLog) * sizeof(U32);
PREFETCH_AREA(dictHashLong, hashTableBytes)
PREFETCH_AREA(dictHashSmall, chainTableBytes)
PREFETCH_AREA(dictHashLong, hashTableBytes);
PREFETCH_AREA(dictHashSmall, chainTableBytes);
}
/* init */
@@ -384,7 +404,7 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
hashLong[h2] = hashSmall[h] = curr; /* update hash tables */
/* check repcode */
if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
if ((ZSTD_index_overlap_check(prefixLowestIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
@@ -393,14 +413,12 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
goto _match_stored;
}
if (matchIndexL > prefixLowestIndex) {
if ((matchIndexL >= prefixLowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
/* check prefix long match */
if (MEM_read64(matchLong) == MEM_read64(ip)) {
mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
offset = (U32)(ip-matchLong);
while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
goto _match_found;
}
mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
offset = (U32)(ip-matchLong);
while (((ip>anchor) & (matchLong>prefixLowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
goto _match_found;
} else if (dictTagsMatchL) {
/* check dictMatchState long match */
U32 const dictMatchIndexL = dictMatchIndexAndTagL >> ZSTD_SHORT_CACHE_TAG_BITS;
@@ -415,7 +433,7 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState_generic(
} }
if (matchIndexS > prefixLowestIndex) {
/* check prefix short match */
/* short match candidate */
if (MEM_read32(match) == MEM_read32(ip)) {
goto _search_next_long;
}
@@ -445,14 +463,12 @@ _search_next_long:
hashLong[hl3] = curr + 1;
/* check prefix long +1 match */
if (matchIndexL3 > prefixLowestIndex) {
if (MEM_read64(matchL3) == MEM_read64(ip+1)) {
mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
ip++;
offset = (U32)(ip-matchL3);
while (((ip>anchor) & (matchL3>prefixLowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
goto _match_found;
}
if ((matchIndexL3 >= prefixLowestIndex) && (MEM_read64(matchL3) == MEM_read64(ip+1))) {
mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
ip++;
offset = (U32)(ip-matchL3);
while (((ip>anchor) & (matchL3>prefixLowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
goto _match_found;
} else if (dictTagsMatchL3) {
/* check dict long +1 match */
U32 const dictMatchIndexL3 = dictMatchIndexAndTagL3 >> ZSTD_SHORT_CACHE_TAG_BITS;
@@ -505,7 +521,7 @@ _match_stored:
const BYTE* repMatch2 = repIndex2 < prefixLowestIndex ?
dictBase + repIndex2 - dictIndexDelta :
base + repIndex2;
if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
if ( (ZSTD_index_overlap_check(prefixLowestIndex, repIndex2))
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixLowest) + 4;
@@ -532,7 +548,7 @@ _match_stored:
#define ZSTD_GEN_DFAST_FN(dictMode, mls) \
static size_t ZSTD_compressBlock_doubleFast_##dictMode##_##mls( \
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], \
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], \
void const* src, size_t srcSize) \
{ \
return ZSTD_compressBlock_doubleFast_##dictMode##_generic(ms, seqStore, rep, src, srcSize, mls); \
@@ -550,7 +566,7 @@ ZSTD_GEN_DFAST_FN(dictMatchState, 7)
size_t ZSTD_compressBlock_doubleFast(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
const U32 mls = ms->cParams.minMatch;
@@ -570,7 +586,7 @@ size_t ZSTD_compressBlock_doubleFast(
size_t ZSTD_compressBlock_doubleFast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
const U32 mls = ms->cParams.minMatch;
@@ -589,8 +605,10 @@ size_t ZSTD_compressBlock_doubleFast_dictMatchState(
}
static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_doubleFast_extDict_generic(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize,
U32 const mls /* template */)
{
@@ -641,7 +659,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
size_t mLength;
hashSmall[hSmall] = hashLong[hLong] = curr; /* update hash table */
if ((((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex doesn't overlap dict + prefix */
if (((ZSTD_index_overlap_check(prefixStartIndex, repIndex))
& (offset_1 <= curr+1 - dictStartIndex)) /* note: we are searching at curr+1 */
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
@@ -709,7 +727,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) /* intentional overflow : ensure repIndex2 doesn't overlap dict + prefix */
if ( ((ZSTD_index_overlap_check(prefixStartIndex, repIndex2))
& (offset_2 <= current2 - dictStartIndex))
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
@@ -739,7 +757,7 @@ ZSTD_GEN_DFAST_FN(extDict, 6)
ZSTD_GEN_DFAST_FN(extDict, 7)
size_t ZSTD_compressBlock_doubleFast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
U32 const mls = ms->cParams.minMatch;
@@ -756,3 +774,5 @@ size_t ZSTD_compressBlock_doubleFast_extDict(
return ZSTD_compressBlock_doubleFast_extDict_7(ms, seqStore, rep, src, srcSize);
}
}
#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */
Utilities/cmzstd/lib/compress/zstd_double_fast.h
+15 -12
View File
@@ -11,29 +11,32 @@
#ifndef ZSTD_DOUBLE_FAST_H
#define ZSTD_DOUBLE_FAST_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "../common/mem.h" /* U32 */
#include "zstd_compress_internal.h" /* ZSTD_CCtx, size_t */
void ZSTD_fillDoubleHashTable(ZSTD_matchState_t* ms,
#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
void ZSTD_fillDoubleHashTable(ZSTD_MatchState_t* ms,
void const* end, ZSTD_dictTableLoadMethod_e dtlm,
ZSTD_tableFillPurpose_e tfp);
size_t ZSTD_compressBlock_doubleFast(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_doubleFast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_doubleFast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#if defined (__cplusplus)
}
#endif
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST ZSTD_compressBlock_doubleFast
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE ZSTD_compressBlock_doubleFast_dictMatchState
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT ZSTD_compressBlock_doubleFast_extDict
#else
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST NULL
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_DOUBLEFAST_EXTDICT NULL
#endif /* ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR */
#endif /* ZSTD_DOUBLE_FAST_H */
Utilities/cmzstd/lib/compress/zstd_fast.c
+103 -78
View File
@@ -11,7 +11,9 @@
#include "zstd_compress_internal.h" /* ZSTD_hashPtr, ZSTD_count, ZSTD_storeSeq */
#include "zstd_fast.h"
static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_fillHashTableForCDict(ZSTD_MatchState_t* ms,
const void* const end,
ZSTD_dictTableLoadMethod_e dtlm)
{
@@ -43,10 +45,12 @@ static void ZSTD_fillHashTableForCDict(ZSTD_matchState_t* ms,
size_t const hashAndTag = ZSTD_hashPtr(ip + p, hBits, mls);
if (hashTable[hashAndTag >> ZSTD_SHORT_CACHE_TAG_BITS] == 0) { /* not yet filled */
ZSTD_writeTaggedIndex(hashTable, hashAndTag, curr + p);
} } } }
} } } }
}
static void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_fillHashTableForCCtx(ZSTD_MatchState_t* ms,
const void* const end,
ZSTD_dictTableLoadMethod_e dtlm)
{
@@ -80,7 +84,7 @@ static void ZSTD_fillHashTableForCCtx(ZSTD_matchState_t* ms,
} } } }
}
void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
void ZSTD_fillHashTable(ZSTD_MatchState_t* ms,
const void* const end,
ZSTD_dictTableLoadMethod_e dtlm,
ZSTD_tableFillPurpose_e tfp)
@@ -93,6 +97,50 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
}
typedef int (*ZSTD_match4Found) (const BYTE* currentPtr, const BYTE* matchAddress, U32 matchIdx, U32 idxLowLimit);
static int
ZSTD_match4Found_cmov(const BYTE* currentPtr, const BYTE* matchAddress, U32 matchIdx, U32 idxLowLimit)
{
/* Array of ~random data, should have low probability of matching data.
* Load from here if the index is invalid.
* Used to avoid unpredictable branches. */
static const BYTE dummy[] = {0x12,0x34,0x56,0x78};
/* currentIdx >= lowLimit is a (somewhat) unpredictable branch.
* However expression below compiles into conditional move.
*/
const BYTE* mvalAddr = ZSTD_selectAddr(matchIdx, idxLowLimit, matchAddress, dummy);
/* Note: this used to be written as : return test1 && test2;
* Unfortunately, once inlined, these tests become branches,
* in which case it becomes critical that they are executed in the right order (test1 then test2).
* So we have to write these tests in a specific manner to ensure their ordering.
*/
if (MEM_read32(currentPtr) != MEM_read32(mvalAddr)) return 0;
/* force ordering of these tests, which matters once the function is inlined, as they become branches */
#if defined(__GNUC__)
__asm__("");
#endif
return matchIdx >= idxLowLimit;
}
static int
ZSTD_match4Found_branch(const BYTE* currentPtr, const BYTE* matchAddress, U32 matchIdx, U32 idxLowLimit)
{
/* using a branch instead of a cmov,
* because it's faster in scenarios where matchIdx >= idxLowLimit is generally true,
* aka almost all candidates are within range */
U32 mval;
if (matchIdx >= idxLowLimit) {
mval = MEM_read32(matchAddress);
} else {
mval = MEM_read32(currentPtr) ^ 1; /* guaranteed to not match. */
}
return (MEM_read32(currentPtr) == mval);
}
/**
* If you squint hard enough (and ignore repcodes), the search operation at any
* given position is broken into 4 stages:
@@ -139,17 +187,17 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
*
* This is also the work we do at the beginning to enter the loop initially.
*/
FORCE_INLINE_TEMPLATE size_t
ZSTD_compressBlock_fast_noDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_fast_noDict_generic(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize,
U32 const mls, U32 const hasStep)
U32 const mls, int useCmov)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable;
U32 const hlog = cParams->hashLog;
/* support stepSize of 0 */
size_t const stepSize = hasStep ? (cParams->targetLength + !(cParams->targetLength) + 1) : 2;
size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1; /* min 2 */
const BYTE* const base = ms->window.base;
const BYTE* const istart = (const BYTE*)src;
const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
@@ -171,8 +219,7 @@ ZSTD_compressBlock_fast_noDict_generic(
size_t hash0; /* hash for ip0 */
size_t hash1; /* hash for ip1 */
U32 idx; /* match idx for ip0 */
U32 mval; /* src value at match idx */
U32 matchIdx; /* match idx for ip0 */
U32 offcode;
const BYTE* match0;
@@ -185,6 +232,7 @@ ZSTD_compressBlock_fast_noDict_generic(
size_t step;
const BYTE* nextStep;
const size_t kStepIncr = (1 << (kSearchStrength - 1));
const ZSTD_match4Found matchFound = useCmov ? ZSTD_match4Found_cmov : ZSTD_match4Found_branch;
DEBUGLOG(5, "ZSTD_compressBlock_fast_generic");
ip0 += (ip0 == prefixStart);
@@ -213,7 +261,7 @@ _start: /* Requires: ip0 */
hash0 = ZSTD_hashPtr(ip0, hlog, mls);
hash1 = ZSTD_hashPtr(ip1, hlog, mls);
idx = hashTable[hash0];
matchIdx = hashTable[hash0];
do {
/* load repcode match for ip[2]*/
@@ -233,35 +281,25 @@ _start: /* Requires: ip0 */
offcode = REPCODE1_TO_OFFBASE;
mLength += 4;
/* First write next hash table entry; we've already calculated it.
* This write is known to be safe because the ip1 is before the
/* Write next hash table entry: it's already calculated.
* This write is known to be safe because ip1 is before the
* repcode (ip2). */
hashTable[hash1] = (U32)(ip1 - base);
goto _match;
}
/* load match for ip[0] */
if (idx >= prefixStartIndex) {
mval = MEM_read32(base + idx);
} else {
mval = MEM_read32(ip0) ^ 1; /* guaranteed to not match. */
}
/* check match at ip[0] */
if (MEM_read32(ip0) == mval) {
/* found a match! */
/* First write next hash table entry; we've already calculated it.
* This write is known to be safe because the ip1 == ip0 + 1, so
* we know we will resume searching after ip1 */
if (matchFound(ip0, base + matchIdx, matchIdx, prefixStartIndex)) {
/* Write next hash table entry (it's already calculated).
* This write is known to be safe because the ip1 == ip0 + 1,
* so searching will resume after ip1 */
hashTable[hash1] = (U32)(ip1 - base);
goto _offset;
}
/* lookup ip[1] */
idx = hashTable[hash1];
matchIdx = hashTable[hash1];
/* hash ip[2] */
hash0 = hash1;
@@ -276,36 +314,19 @@ _start: /* Requires: ip0 */
current0 = (U32)(ip0 - base);
hashTable[hash0] = current0;
/* load match for ip[0] */
if (idx >= prefixStartIndex) {
mval = MEM_read32(base + idx);
} else {
mval = MEM_read32(ip0) ^ 1; /* guaranteed to not match. */
}
/* check match at ip[0] */
if (MEM_read32(ip0) == mval) {
/* found a match! */
/* first write next hash table entry; we've already calculated it */
if (matchFound(ip0, base + matchIdx, matchIdx, prefixStartIndex)) {
/* Write next hash table entry, since it's already calculated */
if (step <= 4) {
/* We need to avoid writing an index into the hash table >= the
* position at which we will pick up our searching after we've
* taken this match.
*
* The minimum possible match has length 4, so the earliest ip0
* can be after we take this match will be the current ip0 + 4.
* ip1 is ip0 + step - 1. If ip1 is >= ip0 + 4, we can't safely
* write this position.
*/
/* Avoid writing an index if it's >= position where search will resume.
* The minimum possible match has length 4, so search can resume at ip0 + 4.
*/
hashTable[hash1] = (U32)(ip1 - base);
}
goto _offset;
}
/* lookup ip[1] */
idx = hashTable[hash1];
matchIdx = hashTable[hash1];
/* hash ip[2] */
hash0 = hash1;
@@ -327,7 +348,7 @@ _start: /* Requires: ip0 */
} while (ip3 < ilimit);
_cleanup:
/* Note that there are probably still a couple positions we could search.
/* Note that there are probably still a couple positions one could search.
* However, it seems to be a meaningful performance hit to try to search
* them. So let's not. */
@@ -356,7 +377,7 @@ _cleanup:
_offset: /* Requires: ip0, idx */
/* Compute the offset code. */
match0 = base + idx;
match0 = base + matchIdx;
rep_offset2 = rep_offset1;
rep_offset1 = (U32)(ip0-match0);
offcode = OFFSET_TO_OFFBASE(rep_offset1);
@@ -401,12 +422,12 @@ _match: /* Requires: ip0, match0, offcode */
goto _start;
}
#define ZSTD_GEN_FAST_FN(dictMode, mls, step) \
static size_t ZSTD_compressBlock_fast_##dictMode##_##mls##_##step( \
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], \
#define ZSTD_GEN_FAST_FN(dictMode, mml, cmov) \
static size_t ZSTD_compressBlock_fast_##dictMode##_##mml##_##cmov( \
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], \
void const* src, size_t srcSize) \
{ \
return ZSTD_compressBlock_fast_##dictMode##_generic(ms, seqStore, rep, src, srcSize, mls, step); \
return ZSTD_compressBlock_fast_##dictMode##_generic(ms, seqStore, rep, src, srcSize, mml, cmov); \
}
ZSTD_GEN_FAST_FN(noDict, 4, 1)
@@ -420,13 +441,15 @@ ZSTD_GEN_FAST_FN(noDict, 6, 0)
ZSTD_GEN_FAST_FN(noDict, 7, 0)
size_t ZSTD_compressBlock_fast(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
U32 const mls = ms->cParams.minMatch;
U32 const mml = ms->cParams.minMatch;
/* use cmov when "candidate in range" branch is likely unpredictable */
int const useCmov = ms->cParams.windowLog < 19;
assert(ms->dictMatchState == NULL);
if (ms->cParams.targetLength > 1) {
switch(mls)
if (useCmov) {
switch(mml)
{
default: /* includes case 3 */
case 4 :
@@ -439,7 +462,8 @@ size_t ZSTD_compressBlock_fast(
return ZSTD_compressBlock_fast_noDict_7_1(ms, seqStore, rep, src, srcSize);
}
} else {
switch(mls)
/* use a branch instead */
switch(mml)
{
default: /* includes case 3 */
case 4 :
@@ -451,13 +475,13 @@ size_t ZSTD_compressBlock_fast(
case 7 :
return ZSTD_compressBlock_fast_noDict_7_0(ms, seqStore, rep, src, srcSize);
}
}
}
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_fast_dictMatchState_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, U32 const mls, U32 const hasStep)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -476,7 +500,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=rep[0], offset_2=rep[1];
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_MatchState_t* const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
const U32* const dictHashTable = dms->hashTable;
const U32 dictStartIndex = dms->window.dictLimit;
@@ -502,7 +526,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
if (ms->prefetchCDictTables) {
size_t const hashTableBytes = (((size_t)1) << dictCParams->hashLog) * sizeof(U32);
PREFETCH_AREA(dictHashTable, hashTableBytes)
PREFETCH_AREA(dictHashTable, hashTableBytes);
}
/* init */
@@ -540,8 +564,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
size_t const dictHashAndTag1 = ZSTD_hashPtr(ip1, dictHBits, mls);
hashTable[hash0] = curr; /* update hash table */
if (((U32) ((prefixStartIndex - 1) - repIndex) >=
3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
if ((ZSTD_index_overlap_check(prefixStartIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip0 + 1))) {
const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip0 + 1 + 4, repMatch + 4, iend, repMatchEnd, prefixStart) + 4;
@@ -574,8 +597,8 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
}
}
if (matchIndex > prefixStartIndex && MEM_read32(match) == MEM_read32(ip0)) {
/* found a regular match */
if (ZSTD_match4Found_cmov(ip0, match, matchIndex, prefixStartIndex)) {
/* found a regular match of size >= 4 */
U32 const offset = (U32) (ip0 - match);
mLength = ZSTD_count(ip0 + 4, match + 4, iend) + 4;
while (((ip0 > anchor) & (match > prefixStart))
@@ -625,7 +648,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
dictBase - dictIndexDelta + repIndex2 :
base + repIndex2;
if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
if ( (ZSTD_index_overlap_check(prefixStartIndex, repIndex2))
&& (MEM_read32(repMatch2) == MEM_read32(ip0))) {
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip0+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
@@ -661,7 +684,7 @@ ZSTD_GEN_FAST_FN(dictMatchState, 6, 0)
ZSTD_GEN_FAST_FN(dictMatchState, 7, 0)
size_t ZSTD_compressBlock_fast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
U32 const mls = ms->cParams.minMatch;
@@ -681,8 +704,10 @@ size_t ZSTD_compressBlock_fast_dictMatchState(
}
static size_t ZSTD_compressBlock_fast_extDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_fast_extDict_generic(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize, U32 const mls, U32 const hasStep)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
@@ -917,7 +942,7 @@ _match: /* Requires: ip0, match0, offcode, matchEnd */
while (ip0 <= ilimit) {
U32 const repIndex2 = (U32)(ip0-base) - offset_2;
const BYTE* const repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (offset_2 > 0)) /* intentional underflow */
if ( ((ZSTD_index_overlap_check(prefixStartIndex, repIndex2)) & (offset_2 > 0))
&& (MEM_read32(repMatch2) == MEM_read32(ip0)) ) {
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip0+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
@@ -940,7 +965,7 @@ ZSTD_GEN_FAST_FN(extDict, 6, 0)
ZSTD_GEN_FAST_FN(extDict, 7, 0)
size_t ZSTD_compressBlock_fast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
U32 const mls = ms->cParams.minMatch;
Utilities/cmzstd/lib/compress/zstd_fast.h
+4 -12
View File
@@ -11,28 +11,20 @@
#ifndef ZSTD_FAST_H
#define ZSTD_FAST_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "../common/mem.h" /* U32 */
#include "zstd_compress_internal.h"
void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
void ZSTD_fillHashTable(ZSTD_MatchState_t* ms,
void const* end, ZSTD_dictTableLoadMethod_e dtlm,
ZSTD_tableFillPurpose_e tfp);
size_t ZSTD_compressBlock_fast(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_fast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_fast_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_FAST_H */
Utilities/cmzstd/lib/compress/zstd_lazy.c
+234 -192
View File
@@ -12,6 +12,11 @@
#include "zstd_lazy.h"
#include "../common/bits.h" /* ZSTD_countTrailingZeros64 */
#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
#define kLazySkippingStep 8
@@ -19,8 +24,9 @@
* Binary Tree search
***************************************/
static void
ZSTD_updateDUBT(ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_updateDUBT(ZSTD_MatchState_t* ms,
const BYTE* ip, const BYTE* iend,
U32 mls)
{
@@ -63,8 +69,9 @@ ZSTD_updateDUBT(ZSTD_matchState_t* ms,
* sort one already inserted but unsorted position
* assumption : curr >= btlow == (curr - btmask)
* doesn't fail */
static void
ZSTD_insertDUBT1(const ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_insertDUBT1(const ZSTD_MatchState_t* ms,
U32 curr, const BYTE* inputEnd,
U32 nbCompares, U32 btLow,
const ZSTD_dictMode_e dictMode)
@@ -152,9 +159,10 @@ ZSTD_insertDUBT1(const ZSTD_matchState_t* ms,
}
static size_t
ZSTD_DUBT_findBetterDictMatch (
const ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_DUBT_findBetterDictMatch (
const ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
size_t* offsetPtr,
size_t bestLength,
@@ -162,7 +170,7 @@ ZSTD_DUBT_findBetterDictMatch (
U32 const mls,
const ZSTD_dictMode_e dictMode)
{
const ZSTD_matchState_t * const dms = ms->dictMatchState;
const ZSTD_MatchState_t * const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dmsCParams = &dms->cParams;
const U32 * const dictHashTable = dms->hashTable;
U32 const hashLog = dmsCParams->hashLog;
@@ -230,8 +238,9 @@ ZSTD_DUBT_findBetterDictMatch (
}
static size_t
ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_DUBT_findBestMatch(ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
size_t* offBasePtr,
U32 const mls,
@@ -381,8 +390,9 @@ ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
FORCE_INLINE_TEMPLATE size_t
ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_BtFindBestMatch( ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iLimit,
size_t* offBasePtr,
const U32 mls /* template */,
@@ -398,7 +408,7 @@ ZSTD_BtFindBestMatch( ZSTD_matchState_t* ms,
* Dedicated dict search
***********************************/
void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip)
void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_MatchState_t* ms, const BYTE* const ip)
{
const BYTE* const base = ms->window.base;
U32 const target = (U32)(ip - base);
@@ -517,7 +527,7 @@ void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const B
*/
FORCE_INLINE_TEMPLATE
size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nbAttempts,
const ZSTD_matchState_t* const dms,
const ZSTD_MatchState_t* const dms,
const BYTE* const ip, const BYTE* const iLimit,
const BYTE* const prefixStart, const U32 curr,
const U32 dictLimit, const size_t ddsIdx) {
@@ -617,8 +627,10 @@ size_t ZSTD_dedicatedDictSearch_lazy_search(size_t* offsetPtr, size_t ml, U32 nb
/* Update chains up to ip (excluded)
Assumption : always within prefix (i.e. not within extDict) */
FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
ZSTD_matchState_t* ms,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_insertAndFindFirstIndex_internal(
ZSTD_MatchState_t* ms,
const ZSTD_compressionParameters* const cParams,
const BYTE* ip, U32 const mls, U32 const lazySkipping)
{
@@ -644,15 +656,16 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_insertAndFindFirstIndex_internal(
return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
}
U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip) {
U32 ZSTD_insertAndFindFirstIndex(ZSTD_MatchState_t* ms, const BYTE* ip) {
const ZSTD_compressionParameters* const cParams = &ms->cParams;
return ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, ms->cParams.minMatch, /* lazySkipping*/ 0);
}
/* inlining is important to hardwire a hot branch (template emulation) */
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_HcFindBestMatch(
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iLimit,
size_t* offsetPtr,
const U32 mls, const ZSTD_dictMode_e dictMode)
@@ -676,7 +689,7 @@ size_t ZSTD_HcFindBestMatch(
U32 nbAttempts = 1U << cParams->searchLog;
size_t ml=4-1;
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_MatchState_t* const dms = ms->dictMatchState;
const U32 ddsHashLog = dictMode == ZSTD_dedicatedDictSearch
? dms->cParams.hashLog - ZSTD_LAZY_DDSS_BUCKET_LOG : 0;
const size_t ddsIdx = dictMode == ZSTD_dedicatedDictSearch
@@ -819,7 +832,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_prefetch(U32 const* hashTable, BYTE const* t
* Fill up the hash cache starting at idx, prefetching up to ZSTD_ROW_HASH_CACHE_SIZE entries,
* but not beyond iLimit.
*/
FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const BYTE* base,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_row_fillHashCache(ZSTD_MatchState_t* ms, const BYTE* base,
U32 const rowLog, U32 const mls,
U32 idx, const BYTE* const iLimit)
{
@@ -845,7 +860,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_fillHashCache(ZSTD_matchState_t* ms, const B
* Returns the hash of base + idx, and replaces the hash in the hash cache with the byte at
* base + idx + ZSTD_ROW_HASH_CACHE_SIZE. Also prefetches the appropriate rows from hashTable and tagTable.
*/
FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTable,
BYTE const* tagTable, BYTE const* base,
U32 idx, U32 const hashLog,
U32 const rowLog, U32 const mls,
@@ -863,10 +880,12 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_row_nextCachedHash(U32* cache, U32 const* hashTab
/* ZSTD_row_update_internalImpl():
* Updates the hash table with positions starting from updateStartIdx until updateEndIdx.
*/
FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
U32 updateStartIdx, U32 const updateEndIdx,
U32 const mls, U32 const rowLog,
U32 const rowMask, U32 const useCache)
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_row_update_internalImpl(ZSTD_MatchState_t* ms,
U32 updateStartIdx, U32 const updateEndIdx,
U32 const mls, U32 const rowLog,
U32 const rowMask, U32 const useCache)
{
U32* const hashTable = ms->hashTable;
BYTE* const tagTable = ms->tagTable;
@@ -892,9 +911,11 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
* Inserts the byte at ip into the appropriate position in the hash table, and updates ms->nextToUpdate.
* Skips sections of long matches as is necessary.
*/
FORCE_INLINE_TEMPLATE void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const BYTE* ip,
U32 const mls, U32 const rowLog,
U32 const rowMask, U32 const useCache)
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_row_update_internal(ZSTD_MatchState_t* ms, const BYTE* ip,
U32 const mls, U32 const rowLog,
U32 const rowMask, U32 const useCache)
{
U32 idx = ms->nextToUpdate;
const BYTE* const base = ms->window.base;
@@ -925,7 +946,7 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internal(ZSTD_matchState_t* ms, const
* External wrapper for ZSTD_row_update_internal(). Used for filling the hashtable during dictionary
* processing.
*/
void ZSTD_row_update(ZSTD_matchState_t* const ms, const BYTE* ip) {
void ZSTD_row_update(ZSTD_MatchState_t* const ms, const BYTE* ip) {
const U32 rowLog = BOUNDED(4, ms->cParams.searchLog, 6);
const U32 rowMask = (1u << rowLog) - 1;
const U32 mls = MIN(ms->cParams.minMatch, 6 /* mls caps out at 6 */);
@@ -1102,22 +1123,23 @@ ZSTD_row_getMatchMask(const BYTE* const tagRow, const BYTE tag, const U32 headGr
/* The high-level approach of the SIMD row based match finder is as follows:
* - Figure out where to insert the new entry:
* - Generate a hash from a byte along with an additional 1-byte "short hash". The additional byte is our "tag"
* - The hashTable is effectively split into groups or "rows" of 16 or 32 entries of U32, and the hash determines
* - Generate a hash for current input position and split it into a one byte of tag and `rowHashLog` bits of index.
* - The hash is salted by a value that changes on every context reset, so when the same table is used
* we will avoid collisions that would otherwise slow us down by introducing phantom matches.
* - The hashTable is effectively split into groups or "rows" of 15 or 31 entries of U32, and the index determines
* which row to insert into.
* - Determine the correct position within the row to insert the entry into. Each row of 16 or 32 can
* be considered as a circular buffer with a "head" index that resides in the tagTable.
* - Also insert the "tag" into the equivalent row and position in the tagTable.
* - Note: The tagTable has 17 or 33 1-byte entries per row, due to 16 or 32 tags, and 1 "head" entry.
* The 17 or 33 entry rows are spaced out to occur every 32 or 64 bytes, respectively,
* for alignment/performance reasons, leaving some bytes unused.
* - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte "short hash" and
* - Determine the correct position within the row to insert the entry into. Each row of 15 or 31 can
* be considered as a circular buffer with a "head" index that resides in the tagTable (overall 16 or 32 bytes
* per row).
* - Use SIMD to efficiently compare the tags in the tagTable to the 1-byte tag calculated for the position and
* generate a bitfield that we can cycle through to check the collisions in the hash table.
* - Pick the longest match.
* - Insert the tag into the equivalent row and position in the tagTable.
*/
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_RowFindBestMatch(
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iLimit,
size_t* offsetPtr,
const U32 mls, const ZSTD_dictMode_e dictMode,
@@ -1149,7 +1171,7 @@ size_t ZSTD_RowFindBestMatch(
U32 hash;
/* DMS/DDS variables that may be referenced laster */
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_MatchState_t* const dms = ms->dictMatchState;
/* Initialize the following variables to satisfy static analyzer */
size_t ddsIdx = 0;
@@ -1318,7 +1340,7 @@ size_t ZSTD_RowFindBestMatch(
* ZSTD_searchMax() dispatches to the correct implementation function.
*
* TODO: The start of the search function involves loading and calculating a
* bunch of constants from the ZSTD_matchState_t. These computations could be
* bunch of constants from the ZSTD_MatchState_t. These computations could be
* done in an initialization function, and saved somewhere in the match state.
* Then we could pass a pointer to the saved state instead of the match state,
* and avoid duplicate computations.
@@ -1342,7 +1364,7 @@ size_t ZSTD_RowFindBestMatch(
#define GEN_ZSTD_BT_SEARCH_FN(dictMode, mls) \
ZSTD_SEARCH_FN_ATTRS size_t ZSTD_BT_SEARCH_FN(dictMode, mls)( \
ZSTD_matchState_t* ms, \
ZSTD_MatchState_t* ms, \
const BYTE* ip, const BYTE* const iLimit, \
size_t* offBasePtr) \
{ \
@@ -1352,7 +1374,7 @@ size_t ZSTD_RowFindBestMatch(
#define GEN_ZSTD_HC_SEARCH_FN(dictMode, mls) \
ZSTD_SEARCH_FN_ATTRS size_t ZSTD_HC_SEARCH_FN(dictMode, mls)( \
ZSTD_matchState_t* ms, \
ZSTD_MatchState_t* ms, \
const BYTE* ip, const BYTE* const iLimit, \
size_t* offsetPtr) \
{ \
@@ -1362,7 +1384,7 @@ size_t ZSTD_RowFindBestMatch(
#define GEN_ZSTD_ROW_SEARCH_FN(dictMode, mls, rowLog) \
ZSTD_SEARCH_FN_ATTRS size_t ZSTD_ROW_SEARCH_FN(dictMode, mls, rowLog)( \
ZSTD_matchState_t* ms, \
ZSTD_MatchState_t* ms, \
const BYTE* ip, const BYTE* const iLimit, \
size_t* offsetPtr) \
{ \
@@ -1463,7 +1485,7 @@ typedef enum { search_hashChain=0, search_binaryTree=1, search_rowHash=2 } searc
* If a match is found its offset is stored in @p offsetPtr.
*/
FORCE_INLINE_TEMPLATE size_t ZSTD_searchMax(
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
const BYTE* ip,
const BYTE* iend,
size_t* offsetPtr,
@@ -1489,9 +1511,10 @@ FORCE_INLINE_TEMPLATE size_t ZSTD_searchMax(
* Common parser - lazy strategy
*********************************/
FORCE_INLINE_TEMPLATE size_t
ZSTD_compressBlock_lazy_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore,
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_lazy_generic(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize,
const searchMethod_e searchMethod, const U32 depth,
@@ -1514,7 +1537,7 @@ ZSTD_compressBlock_lazy_generic(
const int isDMS = dictMode == ZSTD_dictMatchState;
const int isDDS = dictMode == ZSTD_dedicatedDictSearch;
const int isDxS = isDMS || isDDS;
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const ZSTD_MatchState_t* const dms = ms->dictMatchState;
const U32 dictLowestIndex = isDxS ? dms->window.dictLimit : 0;
const BYTE* const dictBase = isDxS ? dms->window.base : NULL;
const BYTE* const dictLowest = isDxS ? dictBase + dictLowestIndex : NULL;
@@ -1567,7 +1590,7 @@ ZSTD_compressBlock_lazy_generic(
&& repIndex < prefixLowestIndex) ?
dictBase + (repIndex - dictIndexDelta) :
base + repIndex;
if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
if ((ZSTD_index_overlap_check(prefixLowestIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
matchLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
@@ -1619,7 +1642,7 @@ ZSTD_compressBlock_lazy_generic(
const BYTE* repMatch = repIndex < prefixLowestIndex ?
dictBase + (repIndex - dictIndexDelta) :
base + repIndex;
if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
if ((ZSTD_index_overlap_check(prefixLowestIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip)) ) {
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
@@ -1655,7 +1678,7 @@ ZSTD_compressBlock_lazy_generic(
const BYTE* repMatch = repIndex < prefixLowestIndex ?
dictBase + (repIndex - dictIndexDelta) :
base + repIndex;
if (((U32)((prefixLowestIndex-1) - repIndex) >= 3 /* intentional underflow */)
if ((ZSTD_index_overlap_check(prefixLowestIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip)) ) {
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
size_t const mlRep = ZSTD_count_2segments(ip+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
@@ -1717,7 +1740,7 @@ _storeSequence:
const BYTE* repMatch = repIndex < prefixLowestIndex ?
dictBase - dictIndexDelta + repIndex :
base + repIndex;
if ( ((U32)((prefixLowestIndex-1) - (U32)repIndex) >= 3 /* intentional overflow */)
if ( (ZSTD_index_overlap_check(prefixLowestIndex, repIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex < prefixLowestIndex ? dictEnd : iend;
matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd2, prefixLowest) + 4;
@@ -1754,154 +1777,165 @@ _storeSequence:
/* Return the last literals size */
return (size_t)(iend - anchor);
}
#endif /* build exclusions */
size_t ZSTD_compressBlock_btlazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
}
#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_greedy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict);
}
size_t ZSTD_compressBlock_btlazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_greedy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dedicatedDictSearch);
}
/* Row-based matchfinder */
size_t ZSTD_compressBlock_lazy2_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict);
}
size_t ZSTD_compressBlock_greedy_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_greedy_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0, ZSTD_dedicatedDictSearch);
}
#endif
#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_lazy_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy_dictMatchState_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1, ZSTD_dedicatedDictSearch);
}
#endif
#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy2(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy2_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dedicatedDictSearch);
}
size_t ZSTD_compressBlock_lazy2_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2, ZSTD_dedicatedDictSearch);
}
#endif
#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btlazy2(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
}
size_t ZSTD_compressBlock_btlazy2_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
}
#endif
#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_compressBlock_lazy_extDict_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore,
ZSTD_MatchState_t* ms, SeqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize,
const searchMethod_e searchMethod, const U32 depth)
@@ -1952,7 +1986,7 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
const U32 repIndex = (U32)(curr+1 - offset_1);
const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
const BYTE* const repMatch = repBase + repIndex;
if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow */
if ( (ZSTD_index_overlap_check(dictLimit, repIndex))
& (offset_1 <= curr+1 - windowLow) ) /* note: we are searching at curr+1 */
if (MEM_read32(ip+1) == MEM_read32(repMatch)) {
/* repcode detected we should take it */
@@ -1993,7 +2027,7 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
const U32 repIndex = (U32)(curr - offset_1);
const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
const BYTE* const repMatch = repBase + repIndex;
if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */
if ( (ZSTD_index_overlap_check(dictLimit, repIndex))
& (offset_1 <= curr - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */
if (MEM_read32(ip) == MEM_read32(repMatch)) {
/* repcode detected */
@@ -2025,7 +2059,7 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
const U32 repIndex = (U32)(curr - offset_1);
const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
const BYTE* const repMatch = repBase + repIndex;
if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */
if ( (ZSTD_index_overlap_check(dictLimit, repIndex))
& (offset_1 <= curr - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */
if (MEM_read32(ip) == MEM_read32(repMatch)) {
/* repcode detected */
@@ -2079,7 +2113,7 @@ _storeSequence:
const U32 repIndex = repCurrent - offset_2;
const BYTE* const repBase = repIndex < dictLimit ? dictBase : base;
const BYTE* const repMatch = repBase + repIndex;
if ( ((U32)((dictLimit-1) - repIndex) >= 3) /* intentional overflow : do not test positions overlapping 2 memory segments */
if ( (ZSTD_index_overlap_check(dictLimit, repIndex))
& (offset_2 <= repCurrent - windowLow) ) /* equivalent to `curr > repIndex >= windowLow` */
if (MEM_read32(ip) == MEM_read32(repMatch)) {
/* repcode detected we should take it */
@@ -2101,57 +2135,65 @@ _storeSequence:
/* Return the last literals size */
return (size_t)(iend - anchor);
}
#endif /* build exclusions */
#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_greedy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0);
}
size_t ZSTD_compressBlock_greedy_extDict_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0);
}
#endif
#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1);
}
size_t ZSTD_compressBlock_lazy_extDict_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1);
}
#endif
#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
}
size_t ZSTD_compressBlock_lazy2_extDict_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2);
}
#endif
#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btlazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
}
size_t ZSTD_compressBlock_greedy_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 0);
}
size_t ZSTD_compressBlock_lazy_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 1);
}
size_t ZSTD_compressBlock_lazy2_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
{
return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_rowHash, 2);
}
#endif
Utilities/cmzstd/lib/compress/zstd_lazy.h
+138 -72
View File
@@ -11,10 +11,6 @@
#ifndef ZSTD_LAZY_H
#define ZSTD_LAZY_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "zstd_compress_internal.h"
/**
@@ -27,101 +23,171 @@ extern "C" {
#define ZSTD_ROW_HASH_TAG_BITS 8 /* nb bits to use for the tag */
U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip);
void ZSTD_row_update(ZSTD_matchState_t* const ms, const BYTE* ip);
#if !defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR)
U32 ZSTD_insertAndFindFirstIndex(ZSTD_MatchState_t* ms, const BYTE* ip);
void ZSTD_row_update(ZSTD_MatchState_t* const ms, const BYTE* ip);
void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_matchState_t* ms, const BYTE* const ip);
void ZSTD_dedicatedDictSearch_lazy_loadDictionary(ZSTD_MatchState_t* ms, const BYTE* const ip);
void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue); /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */
#endif
size_t ZSTD_compressBlock_btlazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#ifndef ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_greedy(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btlazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_dictMatchState_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_dedicatedDictSearch(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_dedicatedDictSearch_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_greedy_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#define ZSTD_COMPRESSBLOCK_GREEDY ZSTD_compressBlock_greedy
#define ZSTD_COMPRESSBLOCK_GREEDY_ROW ZSTD_compressBlock_greedy_row
#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE ZSTD_compressBlock_greedy_dictMatchState
#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW ZSTD_compressBlock_greedy_dictMatchState_row
#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH ZSTD_compressBlock_greedy_dedicatedDictSearch
#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_greedy_dedicatedDictSearch_row
#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT ZSTD_compressBlock_greedy_extDict
#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW ZSTD_compressBlock_greedy_extDict_row
#else
#define ZSTD_COMPRESSBLOCK_GREEDY NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_ROW NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_DICTMATCHSTATE_ROW NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_DEDICATEDDICTSEARCH_ROW NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT NULL
#define ZSTD_COMPRESSBLOCK_GREEDY_EXTDICT_ROW NULL
#endif
#ifndef ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dictMatchState_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_dedicatedDictSearch_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_extDict(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#define ZSTD_COMPRESSBLOCK_LAZY ZSTD_compressBlock_lazy
#define ZSTD_COMPRESSBLOCK_LAZY_ROW ZSTD_compressBlock_lazy_row
#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE ZSTD_compressBlock_lazy_dictMatchState
#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy_dictMatchState_row
#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy_dedicatedDictSearch
#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy_dedicatedDictSearch_row
#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT ZSTD_compressBlock_lazy_extDict
#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW ZSTD_compressBlock_lazy_extDict_row
#else
#define ZSTD_COMPRESSBLOCK_LAZY NULL
#define ZSTD_COMPRESSBLOCK_LAZY_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_LAZY_DICTMATCHSTATE_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH NULL
#define ZSTD_COMPRESSBLOCK_LAZY_DEDICATEDDICTSEARCH_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT NULL
#define ZSTD_COMPRESSBLOCK_LAZY_EXTDICT_ROW NULL
#endif
#ifndef ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_lazy2(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dictMatchState_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_dedicatedDictSearch_row(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_extDict(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_lazy2_extDict_row(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#define ZSTD_COMPRESSBLOCK_LAZY2 ZSTD_compressBlock_lazy2
#define ZSTD_COMPRESSBLOCK_LAZY2_ROW ZSTD_compressBlock_lazy2_row
#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE ZSTD_compressBlock_lazy2_dictMatchState
#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW ZSTD_compressBlock_lazy2_dictMatchState_row
#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH ZSTD_compressBlock_lazy2_dedicatedDictSearch
#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW ZSTD_compressBlock_lazy2_dedicatedDictSearch_row
#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT ZSTD_compressBlock_lazy2_extDict
#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW ZSTD_compressBlock_lazy2_extDict_row
#else
#define ZSTD_COMPRESSBLOCK_LAZY2 NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_DICTMATCHSTATE_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_DEDICATEDDICTSEARCH_ROW NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT NULL
#define ZSTD_COMPRESSBLOCK_LAZY2_EXTDICT_ROW NULL
#endif
#ifndef ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btlazy2(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btlazy2_dictMatchState(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btlazy2_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#if defined (__cplusplus)
}
#define ZSTD_COMPRESSBLOCK_BTLAZY2 ZSTD_compressBlock_btlazy2
#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE ZSTD_compressBlock_btlazy2_dictMatchState
#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT ZSTD_compressBlock_btlazy2_extDict
#else
#define ZSTD_COMPRESSBLOCK_BTLAZY2 NULL
#define ZSTD_COMPRESSBLOCK_BTLAZY2_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_BTLAZY2_EXTDICT NULL
#endif
#endif /* ZSTD_LAZY_H */
Utilities/cmzstd/lib/compress/zstd_ldm.c
+57 -36
View File
@@ -16,7 +16,7 @@
#include "zstd_double_fast.h" /* ZSTD_fillDoubleHashTable() */
#include "zstd_ldm_geartab.h"
#define LDM_BUCKET_SIZE_LOG 3
#define LDM_BUCKET_SIZE_LOG 4
#define LDM_MIN_MATCH_LENGTH 64
#define LDM_HASH_RLOG 7
@@ -133,21 +133,35 @@ done:
}
void ZSTD_ldm_adjustParameters(ldmParams_t* params,
ZSTD_compressionParameters const* cParams)
const ZSTD_compressionParameters* cParams)
{
params->windowLog = cParams->windowLog;
ZSTD_STATIC_ASSERT(LDM_BUCKET_SIZE_LOG <= ZSTD_LDM_BUCKETSIZELOG_MAX);
DEBUGLOG(4, "ZSTD_ldm_adjustParameters");
if (!params->bucketSizeLog) params->bucketSizeLog = LDM_BUCKET_SIZE_LOG;
if (!params->minMatchLength) params->minMatchLength = LDM_MIN_MATCH_LENGTH;
if (params->hashLog == 0) {
params->hashLog = MAX(ZSTD_HASHLOG_MIN, params->windowLog - LDM_HASH_RLOG);
assert(params->hashLog <= ZSTD_HASHLOG_MAX);
}
if (params->hashRateLog == 0) {
params->hashRateLog = params->windowLog < params->hashLog
? 0
: params->windowLog - params->hashLog;
if (params->hashLog > 0) {
/* if params->hashLog is set, derive hashRateLog from it */
assert(params->hashLog <= ZSTD_HASHLOG_MAX);
if (params->windowLog > params->hashLog) {
params->hashRateLog = params->windowLog - params->hashLog;
}
} else {
assert(1 <= (int)cParams->strategy && (int)cParams->strategy <= 9);
/* mapping from [fast, rate7] to [btultra2, rate4] */
params->hashRateLog = 7 - (cParams->strategy/3);
}
}
if (params->hashLog == 0) {
params->hashLog = BOUNDED(ZSTD_HASHLOG_MIN, params->windowLog - params->hashRateLog, ZSTD_HASHLOG_MAX);
}
if (params->minMatchLength == 0) {
params->minMatchLength = LDM_MIN_MATCH_LENGTH;
if (cParams->strategy >= ZSTD_btultra)
params->minMatchLength /= 2;
}
if (params->bucketSizeLog==0) {
assert(1 <= (int)cParams->strategy && (int)cParams->strategy <= 9);
params->bucketSizeLog = BOUNDED(LDM_BUCKET_SIZE_LOG, (U32)cParams->strategy, ZSTD_LDM_BUCKETSIZELOG_MAX);
}
params->bucketSizeLog = MIN(params->bucketSizeLog, params->hashLog);
}
@@ -170,22 +184,22 @@ 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)
const ldmState_t* ldmState, size_t hash, U32 const bucketSizeLog)
{
return ldmState->hashTable + (hash << ldmParams.bucketSizeLog);
return ldmState->hashTable + (hash << bucketSizeLog);
}
/** 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)
U32 const bucketSizeLog)
{
BYTE* const pOffset = ldmState->bucketOffsets + hash;
unsigned const offset = *pOffset;
*(ZSTD_ldm_getBucket(ldmState, hash, ldmParams) + offset) = entry;
*pOffset = (BYTE)((offset + 1) & ((1u << ldmParams.bucketSizeLog) - 1));
*(ZSTD_ldm_getBucket(ldmState, hash, bucketSizeLog) + offset) = entry;
*pOffset = (BYTE)((offset + 1) & ((1u << bucketSizeLog) - 1));
}
@@ -234,7 +248,7 @@ static size_t ZSTD_ldm_countBackwardsMatch_2segments(
*
* The tables for the other strategies are filled within their
* block compressors. */
static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,
static size_t ZSTD_ldm_fillFastTables(ZSTD_MatchState_t* ms,
void const* end)
{
const BYTE* const iend = (const BYTE*)end;
@@ -246,7 +260,11 @@ static size_t ZSTD_ldm_fillFastTables(ZSTD_matchState_t* ms,
break;
case ZSTD_dfast:
#ifndef ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR
ZSTD_fillDoubleHashTable(ms, iend, ZSTD_dtlm_fast, ZSTD_tfp_forCCtx);
#else
assert(0); /* shouldn't be called: cparams should've been adjusted. */
#endif
break;
case ZSTD_greedy:
@@ -269,7 +287,8 @@ void ZSTD_ldm_fillHashTable(
const BYTE* iend, ldmParams_t const* params)
{
U32 const minMatchLength = params->minMatchLength;
U32 const hBits = params->hashLog - params->bucketSizeLog;
U32 const bucketSizeLog = params->bucketSizeLog;
U32 const hBits = params->hashLog - bucketSizeLog;
BYTE const* const base = ldmState->window.base;
BYTE const* const istart = ip;
ldmRollingHashState_t hashState;
@@ -284,7 +303,7 @@ void ZSTD_ldm_fillHashTable(
unsigned n;
numSplits = 0;
hashed = ZSTD_ldm_gear_feed(&hashState, ip, iend - ip, splits, &numSplits);
hashed = ZSTD_ldm_gear_feed(&hashState, ip, (size_t)(iend - ip), splits, &numSplits);
for (n = 0; n < numSplits; n++) {
if (ip + splits[n] >= istart + minMatchLength) {
@@ -295,7 +314,7 @@ void ZSTD_ldm_fillHashTable(
entry.offset = (U32)(split - base);
entry.checksum = (U32)(xxhash >> 32);
ZSTD_ldm_insertEntry(ldmState, hash, entry, *params);
ZSTD_ldm_insertEntry(ldmState, hash, entry, params->bucketSizeLog);
}
}
@@ -309,7 +328,7 @@ void ZSTD_ldm_fillHashTable(
* 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 void ZSTD_ldm_limitTableUpdate(ZSTD_MatchState_t* ms, const BYTE* anchor)
{
U32 const curr = (U32)(anchor - ms->window.base);
if (curr > ms->nextToUpdate + 1024) {
@@ -318,8 +337,10 @@ 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,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_ldm_generateSequences_internal(
ldmState_t* ldmState, RawSeqStore_t* rawSeqStore,
ldmParams_t const* params, void const* src, size_t srcSize)
{
/* LDM parameters */
@@ -373,7 +394,7 @@ static size_t ZSTD_ldm_generateSequences_internal(
candidates[n].split = split;
candidates[n].hash = hash;
candidates[n].checksum = (U32)(xxhash >> 32);
candidates[n].bucket = ZSTD_ldm_getBucket(ldmState, hash, *params);
candidates[n].bucket = ZSTD_ldm_getBucket(ldmState, hash, params->bucketSizeLog);
PREFETCH_L1(candidates[n].bucket);
}
@@ -396,7 +417,7 @@ static size_t ZSTD_ldm_generateSequences_internal(
* the previous one, we merely register it in the hash table and
* move on */
if (split < anchor) {
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, params->bucketSizeLog);
continue;
}
@@ -443,7 +464,7 @@ static size_t ZSTD_ldm_generateSequences_internal(
/* No match found -- insert an entry into the hash table
* and process the next candidate match */
if (bestEntry == NULL) {
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, params->bucketSizeLog);
continue;
}
@@ -464,7 +485,7 @@ static size_t ZSTD_ldm_generateSequences_internal(
/* Insert the current entry into the hash table --- it must be
* done after the previous block to avoid clobbering bestEntry */
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, *params);
ZSTD_ldm_insertEntry(ldmState, hash, newEntry, params->bucketSizeLog);
anchor = split + forwardMatchLength;
@@ -503,7 +524,7 @@ static void ZSTD_ldm_reduceTable(ldmEntry_t* const table, U32 const size,
}
size_t ZSTD_ldm_generateSequences(
ldmState_t* ldmState, rawSeqStore_t* sequences,
ldmState_t* ldmState, RawSeqStore_t* sequences,
ldmParams_t const* params, void const* src, size_t srcSize)
{
U32 const maxDist = 1U << params->windowLog;
@@ -580,7 +601,7 @@ size_t ZSTD_ldm_generateSequences(
}
void
ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize, U32 const minMatch)
ZSTD_ldm_skipSequences(RawSeqStore_t* rawSeqStore, size_t srcSize, U32 const minMatch)
{
while (srcSize > 0 && rawSeqStore->pos < rawSeqStore->size) {
rawSeq* seq = rawSeqStore->seq + rawSeqStore->pos;
@@ -616,7 +637,7 @@ ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize, U32 const min
* 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,
static rawSeq maybeSplitSequence(RawSeqStore_t* rawSeqStore,
U32 const remaining, U32 const minMatch)
{
rawSeq sequence = rawSeqStore->seq[rawSeqStore->pos];
@@ -640,7 +661,7 @@ static rawSeq maybeSplitSequence(rawSeqStore_t* rawSeqStore,
return sequence;
}
void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) {
void ZSTD_ldm_skipRawSeqStoreBytes(RawSeqStore_t* rawSeqStore, size_t nbBytes) {
U32 currPos = (U32)(rawSeqStore->posInSequence + nbBytes);
while (currPos && rawSeqStore->pos < rawSeqStore->size) {
rawSeq currSeq = rawSeqStore->seq[rawSeqStore->pos];
@@ -657,14 +678,14 @@ 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,
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)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
unsigned const minMatch = cParams->minMatch;
ZSTD_blockCompressor const blockCompressor =
ZSTD_BlockCompressor_f const blockCompressor =
ZSTD_selectBlockCompressor(cParams->strategy, useRowMatchFinder, ZSTD_matchState_dictMode(ms));
/* Input bounds */
BYTE const* const istart = (BYTE const*)src;
@@ -689,7 +710,6 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
/* maybeSplitSequence updates rawSeqStore->pos */
rawSeq const sequence = maybeSplitSequence(rawSeqStore,
(U32)(iend - ip), minMatch);
int i;
/* End signal */
if (sequence.offset == 0)
break;
@@ -702,6 +722,7 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
/* Run the block compressor */
DEBUGLOG(5, "pos %u : calling block compressor on segment of size %u", (unsigned)(ip-istart), sequence.litLength);
{
int i;
size_t const newLitLength =
blockCompressor(ms, seqStore, rep, ip, sequence.litLength);
ip += sequence.litLength;
Utilities/cmzstd/lib/compress/zstd_ldm.h
+6 -14
View File
@@ -11,10 +11,6 @@
#ifndef ZSTD_LDM_H
#define ZSTD_LDM_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "zstd_compress_internal.h" /* ldmParams_t, U32 */
#include "../zstd.h" /* ZSTD_CCtx, size_t */
@@ -43,7 +39,7 @@ void ZSTD_ldm_fillHashTable(
* sequences.
*/
size_t ZSTD_ldm_generateSequences(
ldmState_t* ldms, rawSeqStore_t* sequences,
ldmState_t* ldms, RawSeqStore_t* sequences,
ldmParams_t const* params, void const* src, size_t srcSize);
/**
@@ -64,9 +60,9 @@ size_t ZSTD_ldm_generateSequences(
* two. We handle that case correctly, and update `rawSeqStore` appropriately.
* NOTE: This function does not return any errors.
*/
size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_paramSwitch_e useRowMatchFinder,
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);
/**
@@ -76,7 +72,7 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
* Avoids emitting matches less than `minMatch` bytes.
* Must be called for data that is not passed to ZSTD_ldm_blockCompress().
*/
void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize,
void ZSTD_ldm_skipSequences(RawSeqStore_t* rawSeqStore, size_t srcSize,
U32 const minMatch);
/* ZSTD_ldm_skipRawSeqStoreBytes():
@@ -84,7 +80,7 @@ void ZSTD_ldm_skipSequences(rawSeqStore_t* rawSeqStore, size_t srcSize,
* Not to be used in conjunction with ZSTD_ldm_skipSequences().
* Must be called for data with is not passed to ZSTD_ldm_blockCompress().
*/
void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes);
void ZSTD_ldm_skipRawSeqStoreBytes(RawSeqStore_t* rawSeqStore, size_t nbBytes);
/** ZSTD_ldm_getTableSize() :
* Estimate the space needed for long distance matching tables or 0 if LDM is
@@ -110,8 +106,4 @@ size_t ZSTD_ldm_getMaxNbSeq(ldmParams_t params, size_t maxChunkSize);
void ZSTD_ldm_adjustParameters(ldmParams_t* params,
ZSTD_compressionParameters const* cParams);
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_FAST_H */
Utilities/cmzstd/lib/compress/zstd_opt.c
+257 -149
View File
@@ -12,6 +12,9 @@
#include "hist.h"
#include "zstd_opt.h"
#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
#define ZSTD_LITFREQ_ADD 2 /* scaling factor for litFreq, so that frequencies adapt faster to new stats */
#define ZSTD_MAX_PRICE (1<<30)
@@ -264,6 +267,7 @@ static U32 ZSTD_rawLiteralsCost(const BYTE* const literals, U32 const litLength,
const optState_t* const optPtr,
int optLevel)
{
DEBUGLOG(8, "ZSTD_rawLiteralsCost (%u literals)", litLength);
if (litLength == 0) return 0;
if (!ZSTD_compressedLiterals(optPtr))
@@ -402,9 +406,11 @@ MEM_STATIC U32 ZSTD_readMINMATCH(const void* memPtr, U32 length)
/* Update hashTable3 up to ip (excluded)
Assumption : always within prefix (i.e. not within extDict) */
static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms,
U32* nextToUpdate3,
const BYTE* const ip)
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_MatchState_t* ms,
U32* nextToUpdate3,
const BYTE* const ip)
{
U32* const hashTable3 = ms->hashTable3;
U32 const hashLog3 = ms->hashLog3;
@@ -431,8 +437,10 @@ static U32 ZSTD_insertAndFindFirstIndexHash3 (const ZSTD_matchState_t* ms,
* @param ip assumed <= iend-8 .
* @param target The target of ZSTD_updateTree_internal() - we are filling to this position
* @return : nb of positions added */
static U32 ZSTD_insertBt1(
const ZSTD_matchState_t* ms,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_insertBt1(
const ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
U32 const target,
U32 const mls, const int extDict)
@@ -550,15 +558,16 @@ static U32 ZSTD_insertBt1(
}
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_updateTree_internal(
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
const U32 mls, const ZSTD_dictMode_e dictMode)
{
const BYTE* const base = ms->window.base;
U32 const target = (U32)(ip - base);
U32 idx = ms->nextToUpdate;
DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)",
DEBUGLOG(7, "ZSTD_updateTree_internal, from %u to %u (dictMode:%u)",
idx, target, dictMode);
while(idx < target) {
@@ -571,14 +580,16 @@ void ZSTD_updateTree_internal(
ms->nextToUpdate = target;
}
void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend) {
void ZSTD_updateTree(ZSTD_MatchState_t* ms, const BYTE* ip, const BYTE* iend) {
ZSTD_updateTree_internal(ms, ip, iend, ms->cParams.minMatch, ZSTD_noDict);
}
FORCE_INLINE_TEMPLATE U32
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32
ZSTD_insertBtAndGetAllMatches (
ZSTD_match_t* matches, /* store result (found matches) in this table (presumed large enough) */
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
U32* nextToUpdate3,
const BYTE* const ip, const BYTE* const iLimit,
const ZSTD_dictMode_e dictMode,
@@ -614,7 +625,7 @@ ZSTD_insertBtAndGetAllMatches (
U32 mnum = 0;
U32 nbCompares = 1U << cParams->searchLog;
const ZSTD_matchState_t* dms = dictMode == ZSTD_dictMatchState ? ms->dictMatchState : NULL;
const ZSTD_MatchState_t* dms = dictMode == ZSTD_dictMatchState ? ms->dictMatchState : NULL;
const ZSTD_compressionParameters* const dmsCParams =
dictMode == ZSTD_dictMatchState ? &dms->cParams : NULL;
const BYTE* const dmsBase = dictMode == ZSTD_dictMatchState ? dms->window.base : NULL;
@@ -653,13 +664,13 @@ ZSTD_insertBtAndGetAllMatches (
assert(curr >= windowLow);
if ( dictMode == ZSTD_extDict
&& ( ((repOffset-1) /*intentional overflow*/ < curr - windowLow) /* equivalent to `curr > repIndex >= windowLow` */
& (((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */)
& (ZSTD_index_overlap_check(dictLimit, repIndex)) )
&& (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dictEnd, prefixStart) + minMatch;
}
if (dictMode == ZSTD_dictMatchState
&& ( ((repOffset-1) /*intentional overflow*/ < curr - (dmsLowLimit + dmsIndexDelta)) /* equivalent to `curr > repIndex >= dmsLowLimit` */
& ((U32)((dictLimit-1) - repIndex) >= 3) ) /* intentional overflow : do not test positions overlapping 2 memory segments */
& (ZSTD_index_overlap_check(dictLimit, repIndex)) )
&& (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch)) ) {
repLen = (U32)ZSTD_count_2segments(ip+minMatch, repMatch+minMatch, iLimit, dmsEnd, prefixStart) + minMatch;
} }
@@ -808,7 +819,7 @@ ZSTD_insertBtAndGetAllMatches (
typedef U32 (*ZSTD_getAllMatchesFn)(
ZSTD_match_t*,
ZSTD_matchState_t*,
ZSTD_MatchState_t*,
U32*,
const BYTE*,
const BYTE*,
@@ -816,9 +827,11 @@ typedef U32 (*ZSTD_getAllMatchesFn)(
U32 const ll0,
U32 const lengthToBeat);
FORCE_INLINE_TEMPLATE U32 ZSTD_btGetAllMatches_internal(
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
U32 ZSTD_btGetAllMatches_internal(
ZSTD_match_t* matches,
ZSTD_matchState_t* ms,
ZSTD_MatchState_t* ms,
U32* nextToUpdate3,
const BYTE* ip,
const BYTE* const iHighLimit,
@@ -841,7 +854,7 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_btGetAllMatches_internal(
#define GEN_ZSTD_BT_GET_ALL_MATCHES_(dictMode, mls) \
static U32 ZSTD_BT_GET_ALL_MATCHES_FN(dictMode, mls)( \
ZSTD_match_t* matches, \
ZSTD_matchState_t* ms, \
ZSTD_MatchState_t* ms, \
U32* nextToUpdate3, \
const BYTE* ip, \
const BYTE* const iHighLimit, \
@@ -873,7 +886,7 @@ GEN_ZSTD_BT_GET_ALL_MATCHES(dictMatchState)
}
static ZSTD_getAllMatchesFn
ZSTD_selectBtGetAllMatches(ZSTD_matchState_t const* ms, ZSTD_dictMode_e const dictMode)
ZSTD_selectBtGetAllMatches(ZSTD_MatchState_t const* ms, ZSTD_dictMode_e const dictMode)
{
ZSTD_getAllMatchesFn const getAllMatchesFns[3][4] = {
ZSTD_BT_GET_ALL_MATCHES_ARRAY(noDict),
@@ -892,7 +905,7 @@ ZSTD_selectBtGetAllMatches(ZSTD_matchState_t const* ms, ZSTD_dictMode_e const di
/* Struct containing info needed to make decision about ldm inclusion */
typedef struct {
rawSeqStore_t seqStore; /* External match candidates store for this block */
RawSeqStore_t seqStore; /* External match candidates store for this block */
U32 startPosInBlock; /* Start position of the current match candidate */
U32 endPosInBlock; /* End position of the current match candidate */
U32 offset; /* Offset of the match candidate */
@@ -902,7 +915,7 @@ typedef struct {
* Moves forward in @rawSeqStore by @nbBytes,
* which will update the fields 'pos' and 'posInSequence'.
*/
static void ZSTD_optLdm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes)
static void ZSTD_optLdm_skipRawSeqStoreBytes(RawSeqStore_t* rawSeqStore, size_t nbBytes)
{
U32 currPos = (U32)(rawSeqStore->posInSequence + nbBytes);
while (currPos && rawSeqStore->pos < rawSeqStore->size) {
@@ -959,7 +972,7 @@ ZSTD_opt_getNextMatchAndUpdateSeqStore(ZSTD_optLdm_t* optLdm, U32 currPosInBlock
return;
}
/* Matches may be < MINMATCH by this process. In that case, we will reject them
/* Matches may be < minMatch by this process. In that case, we will reject them
when we are deciding whether or not to add the ldm */
optLdm->startPosInBlock = currPosInBlock + literalsBytesRemaining;
optLdm->endPosInBlock = optLdm->startPosInBlock + matchBytesRemaining;
@@ -981,7 +994,8 @@ ZSTD_opt_getNextMatchAndUpdateSeqStore(ZSTD_optLdm_t* optLdm, U32 currPosInBlock
* into 'matches'. Maintains the correct ordering of 'matches'.
*/
static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
const ZSTD_optLdm_t* optLdm, U32 currPosInBlock)
const ZSTD_optLdm_t* optLdm, U32 currPosInBlock,
U32 minMatch)
{
U32 const posDiff = currPosInBlock - optLdm->startPosInBlock;
/* Note: ZSTD_match_t actually contains offBase and matchLength (before subtracting MINMATCH) */
@@ -990,7 +1004,7 @@ static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
/* Ensure that current block position is not outside of the match */
if (currPosInBlock < optLdm->startPosInBlock
|| currPosInBlock >= optLdm->endPosInBlock
|| candidateMatchLength < MINMATCH) {
|| candidateMatchLength < minMatch) {
return;
}
@@ -1010,7 +1024,8 @@ static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
static void
ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm,
ZSTD_match_t* matches, U32* nbMatches,
U32 currPosInBlock, U32 remainingBytes)
U32 currPosInBlock, U32 remainingBytes,
U32 minMatch)
{
if (optLdm->seqStore.size == 0 || optLdm->seqStore.pos >= optLdm->seqStore.size) {
return;
@@ -1027,7 +1042,7 @@ ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm,
}
ZSTD_opt_getNextMatchAndUpdateSeqStore(optLdm, currPosInBlock, remainingBytes);
}
ZSTD_optLdm_maybeAddMatch(matches, nbMatches, optLdm, currPosInBlock);
ZSTD_optLdm_maybeAddMatch(matches, nbMatches, optLdm, currPosInBlock, minMatch);
}
@@ -1035,11 +1050,6 @@ ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm,
* Optimal parser
*********************************/
static U32 ZSTD_totalLen(ZSTD_optimal_t sol)
{
return sol.litlen + sol.mlen;
}
#if 0 /* debug */
static void
@@ -1057,9 +1067,15 @@ listStats(const U32* table, int lastEltID)
#endif
FORCE_INLINE_TEMPLATE size_t
ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
seqStore_t* seqStore,
#define LIT_PRICE(_p) (int)ZSTD_rawLiteralsCost(_p, 1, optStatePtr, optLevel)
#define LL_PRICE(_l) (int)ZSTD_litLengthPrice(_l, optStatePtr, optLevel)
#define LL_INCPRICE(_l) (LL_PRICE(_l) - LL_PRICE(_l-1))
FORCE_INLINE_TEMPLATE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t
ZSTD_compressBlock_opt_generic(ZSTD_MatchState_t* ms,
SeqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize,
const int optLevel,
@@ -1083,10 +1099,10 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
ZSTD_optimal_t* const opt = optStatePtr->priceTable;
ZSTD_match_t* const matches = optStatePtr->matchTable;
ZSTD_optimal_t lastSequence;
ZSTD_optimal_t lastStretch;
ZSTD_optLdm_t optLdm;
ZSTD_memset(&lastSequence, 0, sizeof(ZSTD_optimal_t));
ZSTD_memset(&lastStretch, 0, sizeof(ZSTD_optimal_t));
optLdm.seqStore = ms->ldmSeqStore ? *ms->ldmSeqStore : kNullRawSeqStore;
optLdm.endPosInBlock = optLdm.startPosInBlock = optLdm.offset = 0;
@@ -1108,19 +1124,32 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
U32 const ll0 = !litlen;
U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, ip, iend, rep, ll0, minMatch);
ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
(U32)(ip-istart), (U32)(iend - ip));
if (!nbMatches) { ip++; continue; }
(U32)(ip-istart), (U32)(iend-ip),
minMatch);
if (!nbMatches) {
DEBUGLOG(8, "no match found at cPos %u", (unsigned)(ip-istart));
ip++;
continue;
}
/* Match found: let's store this solution, and eventually find more candidates.
* During this forward pass, @opt is used to store stretches,
* defined as "a match followed by N literals".
* Note how this is different from a Sequence, which is "N literals followed by a match".
* Storing stretches allows us to store different match predecessors
* for each literal position part of a literals run. */
/* initialize opt[0] */
{ U32 i ; for (i=0; i<ZSTD_REP_NUM; i++) opt[0].rep[i] = rep[i]; }
opt[0].mlen = 0; /* means is_a_literal */
opt[0].mlen = 0; /* there are only literals so far */
opt[0].litlen = litlen;
/* We don't need to include the actual price of the literals because
* it is static for the duration of the forward pass, and is included
* in every price. We include the literal length to avoid negative
* prices when we subtract the previous literal length.
/* No need to include the actual price of the literals before the first match
* because it is static for the duration of the forward pass, and is included
* in every subsequent price. But, we include the literal length because
* the cost variation of litlen depends on the value of litlen.
*/
opt[0].price = (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel);
opt[0].price = LL_PRICE(litlen);
ZSTD_STATIC_ASSERT(sizeof(opt[0].rep[0]) == sizeof(rep[0]));
ZSTD_memcpy(&opt[0].rep, rep, sizeof(opt[0].rep));
/* large match -> immediate encoding */
{ U32 const maxML = matches[nbMatches-1].len;
@@ -1129,82 +1158,106 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
nbMatches, maxML, maxOffBase, (U32)(ip-prefixStart));
if (maxML > sufficient_len) {
lastSequence.litlen = litlen;
lastSequence.mlen = maxML;
lastSequence.off = maxOffBase;
DEBUGLOG(6, "large match (%u>%u), immediate encoding",
lastStretch.litlen = 0;
lastStretch.mlen = maxML;
lastStretch.off = maxOffBase;
DEBUGLOG(6, "large match (%u>%u) => immediate encoding",
maxML, sufficient_len);
cur = 0;
last_pos = ZSTD_totalLen(lastSequence);
last_pos = maxML;
goto _shortestPath;
} }
/* set prices for first matches starting position == 0 */
assert(opt[0].price >= 0);
{ U32 const literalsPrice = (U32)opt[0].price + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
U32 pos;
{ U32 pos;
U32 matchNb;
for (pos = 1; pos < minMatch; pos++) {
opt[pos].price = ZSTD_MAX_PRICE; /* mlen, litlen and price will be fixed during forward scanning */
opt[pos].price = ZSTD_MAX_PRICE;
opt[pos].mlen = 0;
opt[pos].litlen = litlen + pos;
}
for (matchNb = 0; matchNb < nbMatches; matchNb++) {
U32 const offBase = matches[matchNb].off;
U32 const end = matches[matchNb].len;
for ( ; pos <= end ; pos++ ) {
U32 const matchPrice = ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel);
U32 const sequencePrice = literalsPrice + matchPrice;
int const matchPrice = (int)ZSTD_getMatchPrice(offBase, pos, optStatePtr, optLevel);
int const sequencePrice = opt[0].price + matchPrice;
DEBUGLOG(7, "rPos:%u => set initial price : %.2f",
pos, ZSTD_fCost((int)sequencePrice));
pos, ZSTD_fCost(sequencePrice));
opt[pos].mlen = pos;
opt[pos].off = offBase;
opt[pos].litlen = litlen;
opt[pos].price = (int)sequencePrice;
} }
opt[pos].litlen = 0; /* end of match */
opt[pos].price = sequencePrice + LL_PRICE(0);
}
}
last_pos = pos-1;
opt[pos].price = ZSTD_MAX_PRICE;
}
}
/* check further positions */
for (cur = 1; cur <= last_pos; cur++) {
const BYTE* const inr = ip + cur;
assert(cur < ZSTD_OPT_NUM);
DEBUGLOG(7, "cPos:%zi==rPos:%u", inr-istart, cur)
assert(cur <= ZSTD_OPT_NUM);
DEBUGLOG(7, "cPos:%i==rPos:%u", (int)(inr-istart), cur);
/* Fix current position with one literal if cheaper */
{ U32 const litlen = (opt[cur-1].mlen == 0) ? opt[cur-1].litlen + 1 : 1;
{ U32 const litlen = opt[cur-1].litlen + 1;
int const price = opt[cur-1].price
+ (int)ZSTD_rawLiteralsCost(ip+cur-1, 1, optStatePtr, optLevel)
+ (int)ZSTD_litLengthPrice(litlen, optStatePtr, optLevel)
- (int)ZSTD_litLengthPrice(litlen-1, optStatePtr, optLevel);
+ LIT_PRICE(ip+cur-1)
+ LL_INCPRICE(litlen);
assert(price < 1000000000); /* overflow check */
if (price <= opt[cur].price) {
DEBUGLOG(7, "cPos:%zi==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)",
inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen,
ZSTD_optimal_t const prevMatch = opt[cur];
DEBUGLOG(7, "cPos:%i==rPos:%u : better price (%.2f<=%.2f) using literal (ll==%u) (hist:%u,%u,%u)",
(int)(inr-istart), cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price), litlen,
opt[cur-1].rep[0], opt[cur-1].rep[1], opt[cur-1].rep[2]);
opt[cur].mlen = 0;
opt[cur].off = 0;
opt[cur] = opt[cur-1];
opt[cur].litlen = litlen;
opt[cur].price = price;
if ( (optLevel >= 1) /* additional check only for higher modes */
&& (prevMatch.litlen == 0) /* replace a match */
&& (LL_INCPRICE(1) < 0) /* ll1 is cheaper than ll0 */
&& LIKELY(ip + cur < iend)
) {
/* check next position, in case it would be cheaper */
int with1literal = prevMatch.price + LIT_PRICE(ip+cur) + LL_INCPRICE(1);
int withMoreLiterals = price + LIT_PRICE(ip+cur) + LL_INCPRICE(litlen+1);
DEBUGLOG(7, "then at next rPos %u : match+1lit %.2f vs %ulits %.2f",
cur+1, ZSTD_fCost(with1literal), litlen+1, ZSTD_fCost(withMoreLiterals));
if ( (with1literal < withMoreLiterals)
&& (with1literal < opt[cur+1].price) ) {
/* update offset history - before it disappears */
U32 const prev = cur - prevMatch.mlen;
Repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, prevMatch.off, opt[prev].litlen==0);
assert(cur >= prevMatch.mlen);
DEBUGLOG(7, "==> match+1lit is cheaper (%.2f < %.2f) (hist:%u,%u,%u) !",
ZSTD_fCost(with1literal), ZSTD_fCost(withMoreLiterals),
newReps.rep[0], newReps.rep[1], newReps.rep[2] );
opt[cur+1] = prevMatch; /* mlen & offbase */
ZSTD_memcpy(opt[cur+1].rep, &newReps, sizeof(Repcodes_t));
opt[cur+1].litlen = 1;
opt[cur+1].price = with1literal;
if (last_pos < cur+1) last_pos = cur+1;
}
}
} else {
DEBUGLOG(7, "cPos:%zi==rPos:%u : literal would cost more (%.2f>%.2f) (hist:%u,%u,%u)",
inr-istart, cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price),
opt[cur].rep[0], opt[cur].rep[1], opt[cur].rep[2]);
DEBUGLOG(7, "cPos:%i==rPos:%u : literal would cost more (%.2f>%.2f)",
(int)(inr-istart), cur, ZSTD_fCost(price), ZSTD_fCost(opt[cur].price));
}
}
/* Set the repcodes of the current position. We must do it here
* because we rely on the repcodes of the 2nd to last sequence being
* correct to set the next chunks repcodes during the backward
* traversal.
/* Offset history is not updated during match comparison.
* Do it here, now that the match is selected and confirmed.
*/
ZSTD_STATIC_ASSERT(sizeof(opt[cur].rep) == sizeof(repcodes_t));
ZSTD_STATIC_ASSERT(sizeof(opt[cur].rep) == sizeof(Repcodes_t));
assert(cur >= opt[cur].mlen);
if (opt[cur].mlen != 0) {
if (opt[cur].litlen == 0) {
/* just finished a match => alter offset history */
U32 const prev = cur - opt[cur].mlen;
repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[cur].litlen==0);
ZSTD_memcpy(opt[cur].rep, &newReps, sizeof(repcodes_t));
} else {
ZSTD_memcpy(opt[cur].rep, opt[cur - 1].rep, sizeof(repcodes_t));
Repcodes_t const newReps = ZSTD_newRep(opt[prev].rep, opt[cur].off, opt[prev].litlen==0);
ZSTD_memcpy(opt[cur].rep, &newReps, sizeof(Repcodes_t));
}
/* last match must start at a minimum distance of 8 from oend */
@@ -1214,38 +1267,37 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
if ( (optLevel==0) /*static_test*/
&& (opt[cur+1].price <= opt[cur].price + (BITCOST_MULTIPLIER/2)) ) {
DEBUGLOG(7, "move to next rPos:%u : price is <=", cur+1);
DEBUGLOG(7, "skip current position : next rPos(%u) price is cheaper", cur+1);
continue; /* skip unpromising positions; about ~+6% speed, -0.01 ratio */
}
assert(opt[cur].price >= 0);
{ U32 const ll0 = (opt[cur].mlen != 0);
U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;
U32 const previousPrice = (U32)opt[cur].price;
U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
{ U32 const ll0 = (opt[cur].litlen == 0);
int const previousPrice = opt[cur].price;
int const basePrice = previousPrice + LL_PRICE(0);
U32 nbMatches = getAllMatches(matches, ms, &nextToUpdate3, inr, iend, opt[cur].rep, ll0, minMatch);
U32 matchNb;
ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
(U32)(inr-istart), (U32)(iend-inr));
(U32)(inr-istart), (U32)(iend-inr),
minMatch);
if (!nbMatches) {
DEBUGLOG(7, "rPos:%u : no match found", cur);
continue;
}
{ U32 const maxML = matches[nbMatches-1].len;
DEBUGLOG(7, "cPos:%zi==rPos:%u, found %u matches, of maxLength=%u",
inr-istart, cur, nbMatches, maxML);
{ U32 const longestML = matches[nbMatches-1].len;
DEBUGLOG(7, "cPos:%i==rPos:%u, found %u matches, of longest ML=%u",
(int)(inr-istart), cur, nbMatches, longestML);
if ( (maxML > sufficient_len)
|| (cur + maxML >= ZSTD_OPT_NUM) ) {
lastSequence.mlen = maxML;
lastSequence.off = matches[nbMatches-1].off;
lastSequence.litlen = litlen;
cur -= (opt[cur].mlen==0) ? opt[cur].litlen : 0; /* last sequence is actually only literals, fix cur to last match - note : may underflow, in which case, it's first sequence, and it's okay */
last_pos = cur + ZSTD_totalLen(lastSequence);
if (cur > ZSTD_OPT_NUM) cur = 0; /* underflow => first match */
if ( (longestML > sufficient_len)
|| (cur + longestML >= ZSTD_OPT_NUM)
|| (ip + cur + longestML >= iend) ) {
lastStretch.mlen = longestML;
lastStretch.off = matches[nbMatches-1].off;
lastStretch.litlen = 0;
last_pos = cur + longestML;
goto _shortestPath;
} }
@@ -1257,19 +1309,24 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
U32 mlen;
DEBUGLOG(7, "testing match %u => offBase=%4u, mlen=%2u, llen=%2u",
matchNb, matches[matchNb].off, lastML, litlen);
matchNb, matches[matchNb].off, lastML, opt[cur].litlen);
for (mlen = lastML; mlen >= startML; mlen--) { /* scan downward */
U32 const pos = cur + mlen;
int const price = (int)basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);
int const price = basePrice + (int)ZSTD_getMatchPrice(offset, mlen, optStatePtr, optLevel);
if ((pos > last_pos) || (price < opt[pos].price)) {
DEBUGLOG(7, "rPos:%u (ml=%2u) => new better price (%.2f<%.2f)",
pos, mlen, ZSTD_fCost(price), ZSTD_fCost(opt[pos].price));
while (last_pos < pos) { opt[last_pos+1].price = ZSTD_MAX_PRICE; last_pos++; } /* fill empty positions */
while (last_pos < pos) {
/* fill empty positions, for future comparisons */
last_pos++;
opt[last_pos].price = ZSTD_MAX_PRICE;
opt[last_pos].litlen = !0; /* just needs to be != 0, to mean "not an end of match" */
}
opt[pos].mlen = mlen;
opt[pos].off = offset;
opt[pos].litlen = litlen;
opt[pos].litlen = 0;
opt[pos].price = price;
} else {
DEBUGLOG(7, "rPos:%u (ml=%2u) => new price is worse (%.2f>=%.2f)",
@@ -1277,55 +1334,89 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
if (optLevel==0) break; /* early update abort; gets ~+10% speed for about -0.01 ratio loss */
}
} } }
opt[last_pos+1].price = ZSTD_MAX_PRICE;
} /* for (cur = 1; cur <= last_pos; cur++) */
lastSequence = opt[last_pos];
cur = last_pos > ZSTD_totalLen(lastSequence) ? last_pos - ZSTD_totalLen(lastSequence) : 0; /* single sequence, and it starts before `ip` */
assert(cur < ZSTD_OPT_NUM); /* control overflow*/
lastStretch = opt[last_pos];
assert(cur >= lastStretch.mlen);
cur = last_pos - lastStretch.mlen;
_shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */
assert(opt[0].mlen == 0);
assert(last_pos >= lastStretch.mlen);
assert(cur == last_pos - lastStretch.mlen);
/* Set the next chunk's repcodes based on the repcodes of the beginning
* of the last match, and the last sequence. This avoids us having to
* update them while traversing the sequences.
*/
if (lastSequence.mlen != 0) {
repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastSequence.off, lastSequence.litlen==0);
ZSTD_memcpy(rep, &reps, sizeof(reps));
if (lastStretch.mlen==0) {
/* no solution : all matches have been converted into literals */
assert(lastStretch.litlen == (ip - anchor) + last_pos);
ip += last_pos;
continue;
}
assert(lastStretch.off > 0);
/* Update offset history */
if (lastStretch.litlen == 0) {
/* finishing on a match : update offset history */
Repcodes_t const reps = ZSTD_newRep(opt[cur].rep, lastStretch.off, opt[cur].litlen==0);
ZSTD_memcpy(rep, &reps, sizeof(Repcodes_t));
} else {
ZSTD_memcpy(rep, opt[cur].rep, sizeof(repcodes_t));
ZSTD_memcpy(rep, lastStretch.rep, sizeof(Repcodes_t));
assert(cur >= lastStretch.litlen);
cur -= lastStretch.litlen;
}
{ U32 const storeEnd = cur + 1;
/* Let's write the shortest path solution.
* It is stored in @opt in reverse order,
* starting from @storeEnd (==cur+2),
* effectively partially @opt overwriting.
* Content is changed too:
* - So far, @opt stored stretches, aka a match followed by literals
* - Now, it will store sequences, aka literals followed by a match
*/
{ U32 const storeEnd = cur + 2;
U32 storeStart = storeEnd;
U32 seqPos = cur;
U32 stretchPos = cur;
DEBUGLOG(6, "start reverse traversal (last_pos:%u, cur:%u)",
last_pos, cur); (void)last_pos;
assert(storeEnd < ZSTD_OPT_NUM);
DEBUGLOG(6, "last sequence copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
storeEnd, lastSequence.litlen, lastSequence.mlen, lastSequence.off);
opt[storeEnd] = lastSequence;
while (seqPos > 0) {
U32 const backDist = ZSTD_totalLen(opt[seqPos]);
assert(storeEnd < ZSTD_OPT_SIZE);
DEBUGLOG(6, "last stretch copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
storeEnd, lastStretch.litlen, lastStretch.mlen, lastStretch.off);
if (lastStretch.litlen > 0) {
/* last "sequence" is unfinished: just a bunch of literals */
opt[storeEnd].litlen = lastStretch.litlen;
opt[storeEnd].mlen = 0;
storeStart = storeEnd-1;
opt[storeStart] = lastStretch;
} {
opt[storeEnd] = lastStretch; /* note: litlen will be fixed */
storeStart = storeEnd;
}
while (1) {
ZSTD_optimal_t nextStretch = opt[stretchPos];
opt[storeStart].litlen = nextStretch.litlen;
DEBUGLOG(6, "selected sequence (llen=%u,mlen=%u,ofc=%u)",
opt[storeStart].litlen, opt[storeStart].mlen, opt[storeStart].off);
if (nextStretch.mlen == 0) {
/* reaching beginning of segment */
break;
}
storeStart--;
DEBUGLOG(6, "sequence from rPos=%u copied into pos=%u (llen=%u,mlen=%u,ofc=%u)",
seqPos, storeStart, opt[seqPos].litlen, opt[seqPos].mlen, opt[seqPos].off);
opt[storeStart] = opt[seqPos];
seqPos = (seqPos > backDist) ? seqPos - backDist : 0;
opt[storeStart] = nextStretch; /* note: litlen will be fixed */
assert(nextStretch.litlen + nextStretch.mlen <= stretchPos);
stretchPos -= nextStretch.litlen + nextStretch.mlen;
}
/* save sequences */
DEBUGLOG(6, "sending selected sequences into seqStore")
DEBUGLOG(6, "sending selected sequences into seqStore");
{ U32 storePos;
for (storePos=storeStart; storePos <= storeEnd; storePos++) {
U32 const llen = opt[storePos].litlen;
U32 const mlen = opt[storePos].mlen;
U32 const offBase = opt[storePos].off;
U32 const advance = llen + mlen;
DEBUGLOG(6, "considering seq starting at %zi, llen=%u, mlen=%u",
anchor - istart, (unsigned)llen, (unsigned)mlen);
DEBUGLOG(6, "considering seq starting at %i, llen=%u, mlen=%u",
(int)(anchor - istart), (unsigned)llen, (unsigned)mlen);
if (mlen==0) { /* only literals => must be last "sequence", actually starting a new stream of sequences */
assert(storePos == storeEnd); /* must be last sequence */
@@ -1339,6 +1430,9 @@ _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */
anchor += advance;
ip = anchor;
} }
DEBUGLOG(7, "new offset history : %u, %u, %u", rep[0], rep[1], rep[2]);
/* update all costs */
ZSTD_setBasePrices(optStatePtr, optLevel);
}
} /* while (ip < ilimit) */
@@ -1346,42 +1440,51 @@ _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */
/* Return the last literals size */
return (size_t)(iend - anchor);
}
#endif /* build exclusions */
#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
static size_t ZSTD_compressBlock_opt0(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode)
{
return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 0 /* optLevel */, dictMode);
}
#endif
#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
static size_t ZSTD_compressBlock_opt2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize, const ZSTD_dictMode_e dictMode)
{
return ZSTD_compressBlock_opt_generic(ms, seqStore, rep, src, srcSize, 2 /* optLevel */, dictMode);
}
#endif
#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btopt(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
DEBUGLOG(5, "ZSTD_compressBlock_btopt");
return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_noDict);
}
#endif
#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
/* ZSTD_initStats_ultra():
* make a first compression pass, just to seed stats with more accurate starting values.
* only works on first block, with no dictionary and no ldm.
* this function cannot error out, its narrow contract must be respected.
*/
static void
ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
seqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_initStats_ultra(ZSTD_MatchState_t* ms,
SeqStore_t* seqStore,
U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
U32 tmpRep[ZSTD_REP_NUM]; /* updated rep codes will sink here */
ZSTD_memcpy(tmpRep, rep, sizeof(tmpRep));
@@ -1404,7 +1507,7 @@ ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
}
size_t ZSTD_compressBlock_btultra(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
DEBUGLOG(5, "ZSTD_compressBlock_btultra (srcSize=%zu)", srcSize);
@@ -1412,7 +1515,7 @@ size_t ZSTD_compressBlock_btultra(
}
size_t ZSTD_compressBlock_btultra2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
U32 const curr = (U32)((const BYTE*)src - ms->window.base);
@@ -1425,7 +1528,7 @@ size_t ZSTD_compressBlock_btultra2(
* Consequently, this can only work if no data has been previously loaded in tables,
* aka, no dictionary, no prefix, no ldm preprocessing.
* The compression ratio gain is generally small (~0.5% on first block),
** the cost is 2x cpu time on first block. */
* the cost is 2x cpu time on first block. */
assert(srcSize <= ZSTD_BLOCKSIZE_MAX);
if ( (ms->opt.litLengthSum==0) /* first block */
&& (seqStore->sequences == seqStore->sequencesStart) /* no ldm */
@@ -1438,34 +1541,39 @@ size_t ZSTD_compressBlock_btultra2(
return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_noDict);
}
#endif
#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btopt_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_btopt_extDict(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_extDict);
}
#endif
#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btultra_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_dictMatchState);
}
size_t ZSTD_compressBlock_btopt_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
return ZSTD_compressBlock_opt0(ms, seqStore, rep, src, srcSize, ZSTD_extDict);
}
size_t ZSTD_compressBlock_btultra_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
const void* src, size_t srcSize)
{
return ZSTD_compressBlock_opt2(ms, seqStore, rep, src, srcSize, ZSTD_extDict);
}
#endif
/* note : no btultra2 variant for extDict nor dictMatchState,
* because btultra2 is not meant to work with dictionaries
Utilities/cmzstd/lib/compress/zstd_opt.h
+39 -23
View File
@@ -11,46 +11,62 @@
#ifndef ZSTD_OPT_H
#define ZSTD_OPT_H
#if defined (__cplusplus)
extern "C" {
#endif
#include "zstd_compress_internal.h"
#if !defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
|| !defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
/* used in ZSTD_loadDictionaryContent() */
void ZSTD_updateTree(ZSTD_matchState_t* ms, const BYTE* ip, const BYTE* iend);
void ZSTD_updateTree(ZSTD_MatchState_t* ms, const BYTE* ip, const BYTE* iend);
#endif
#ifndef ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btopt(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btultra(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btultra2(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btopt_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btopt_extDict(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#define ZSTD_COMPRESSBLOCK_BTOPT ZSTD_compressBlock_btopt
#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE ZSTD_compressBlock_btopt_dictMatchState
#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT ZSTD_compressBlock_btopt_extDict
#else
#define ZSTD_COMPRESSBLOCK_BTOPT NULL
#define ZSTD_COMPRESSBLOCK_BTOPT_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_BTOPT_EXTDICT NULL
#endif
#ifndef ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR
size_t ZSTD_compressBlock_btultra(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btultra_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btopt_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
size_t ZSTD_compressBlock_btultra_extDict(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
/* note : no btultra2 variant for extDict nor dictMatchState,
* because btultra2 is not meant to work with dictionaries
* and is only specific for the first block (no prefix) */
size_t ZSTD_compressBlock_btultra2(
ZSTD_MatchState_t* ms, SeqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize);
#if defined (__cplusplus)
}
#define ZSTD_COMPRESSBLOCK_BTULTRA ZSTD_compressBlock_btultra
#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE ZSTD_compressBlock_btultra_dictMatchState
#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT ZSTD_compressBlock_btultra_extDict
#define ZSTD_COMPRESSBLOCK_BTULTRA2 ZSTD_compressBlock_btultra2
#else
#define ZSTD_COMPRESSBLOCK_BTULTRA NULL
#define ZSTD_COMPRESSBLOCK_BTULTRA_DICTMATCHSTATE NULL
#define ZSTD_COMPRESSBLOCK_BTULTRA_EXTDICT NULL
#define ZSTD_COMPRESSBLOCK_BTULTRA2 NULL
#endif
#endif /* ZSTD_OPT_H */
Utilities/cmzstd/lib/compress/zstd_preSplit.c
+238
View File
@@ -0,0 +1,238 @@
/*
* Copyright (c) Meta Platforms, Inc. and affiliates.
* 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 "../common/compiler.h" /* ZSTD_ALIGNOF */
#include "../common/mem.h" /* S64 */
#include "../common/zstd_deps.h" /* ZSTD_memset */
#include "../common/zstd_internal.h" /* ZSTD_STATIC_ASSERT */
#include "hist.h" /* HIST_add */
#include "zstd_preSplit.h"
#define BLOCKSIZE_MIN 3500
#define THRESHOLD_PENALTY_RATE 16
#define THRESHOLD_BASE (THRESHOLD_PENALTY_RATE - 2)
#define THRESHOLD_PENALTY 3
#define HASHLENGTH 2
#define HASHLOG_MAX 10
#define HASHTABLESIZE (1 << HASHLOG_MAX)
#define HASHMASK (HASHTABLESIZE - 1)
#define KNUTH 0x9e3779b9
/* for hashLog > 8, hash 2 bytes.
* for hashLog == 8, just take the byte, no hashing.
* The speed of this method relies on compile-time constant propagation */
FORCE_INLINE_TEMPLATE unsigned hash2(const void *p, unsigned hashLog)
{
assert(hashLog >= 8);
if (hashLog == 8) return (U32)((const BYTE*)p)[0];
assert(hashLog <= HASHLOG_MAX);
return (U32)(MEM_read16(p)) * KNUTH >> (32 - hashLog);
}
typedef struct {
unsigned events[HASHTABLESIZE];
size_t nbEvents;
} Fingerprint;
typedef struct {
Fingerprint pastEvents;
Fingerprint newEvents;
} FPStats;
static void initStats(FPStats* fpstats)
{
ZSTD_memset(fpstats, 0, sizeof(FPStats));
}
FORCE_INLINE_TEMPLATE void
addEvents_generic(Fingerprint* fp, const void* src, size_t srcSize, size_t samplingRate, unsigned hashLog)
{
const char* p = (const char*)src;
size_t limit = srcSize - HASHLENGTH + 1;
size_t n;
assert(srcSize >= HASHLENGTH);
for (n = 0; n < limit; n+=samplingRate) {
fp->events[hash2(p+n, hashLog)]++;
}
fp->nbEvents += limit/samplingRate;
}
FORCE_INLINE_TEMPLATE void
recordFingerprint_generic(Fingerprint* fp, const void* src, size_t srcSize, size_t samplingRate, unsigned hashLog)
{
ZSTD_memset(fp, 0, sizeof(unsigned) * ((size_t)1 << hashLog));
fp->nbEvents = 0;
addEvents_generic(fp, src, srcSize, samplingRate, hashLog);
}
typedef void (*RecordEvents_f)(Fingerprint* fp, const void* src, size_t srcSize);
#define FP_RECORD(_rate) ZSTD_recordFingerprint_##_rate
#define ZSTD_GEN_RECORD_FINGERPRINT(_rate, _hSize) \
static void FP_RECORD(_rate)(Fingerprint* fp, const void* src, size_t srcSize) \
{ \
recordFingerprint_generic(fp, src, srcSize, _rate, _hSize); \
}
ZSTD_GEN_RECORD_FINGERPRINT(1, 10)
ZSTD_GEN_RECORD_FINGERPRINT(5, 10)
ZSTD_GEN_RECORD_FINGERPRINT(11, 9)
ZSTD_GEN_RECORD_FINGERPRINT(43, 8)
static U64 abs64(S64 s64) { return (U64)((s64 < 0) ? -s64 : s64); }
static U64 fpDistance(const Fingerprint* fp1, const Fingerprint* fp2, unsigned hashLog)
{
U64 distance = 0;
size_t n;
assert(hashLog <= HASHLOG_MAX);
for (n = 0; n < ((size_t)1 << hashLog); n++) {
distance +=
abs64((S64)fp1->events[n] * (S64)fp2->nbEvents - (S64)fp2->events[n] * (S64)fp1->nbEvents);
}
return distance;
}
/* Compare newEvents with pastEvents
* return 1 when considered "too different"
*/
static int compareFingerprints(const Fingerprint* ref,
const Fingerprint* newfp,
int penalty,
unsigned hashLog)
{
assert(ref->nbEvents > 0);
assert(newfp->nbEvents > 0);
{ U64 p50 = (U64)ref->nbEvents * (U64)newfp->nbEvents;
U64 deviation = fpDistance(ref, newfp, hashLog);
U64 threshold = p50 * (U64)(THRESHOLD_BASE + penalty) / THRESHOLD_PENALTY_RATE;
return deviation >= threshold;
}
}
static void mergeEvents(Fingerprint* acc, const Fingerprint* newfp)
{
size_t n;
for (n = 0; n < HASHTABLESIZE; n++) {
acc->events[n] += newfp->events[n];
}
acc->nbEvents += newfp->nbEvents;
}
static void flushEvents(FPStats* fpstats)
{
size_t n;
for (n = 0; n < HASHTABLESIZE; n++) {
fpstats->pastEvents.events[n] = fpstats->newEvents.events[n];
}
fpstats->pastEvents.nbEvents = fpstats->newEvents.nbEvents;
ZSTD_memset(&fpstats->newEvents, 0, sizeof(fpstats->newEvents));
}
static void removeEvents(Fingerprint* acc, const Fingerprint* slice)
{
size_t n;
for (n = 0; n < HASHTABLESIZE; n++) {
assert(acc->events[n] >= slice->events[n]);
acc->events[n] -= slice->events[n];
}
acc->nbEvents -= slice->nbEvents;
}
#define CHUNKSIZE (8 << 10)
static size_t ZSTD_splitBlock_byChunks(const void* blockStart, size_t blockSize,
int level,
void* workspace, size_t wkspSize)
{
static const RecordEvents_f records_fs[] = {
FP_RECORD(43), FP_RECORD(11), FP_RECORD(5), FP_RECORD(1)
};
static const unsigned hashParams[] = { 8, 9, 10, 10 };
const RecordEvents_f record_f = (assert(0<=level && level<=3), records_fs[level]);
FPStats* const fpstats = (FPStats*)workspace;
const char* p = (const char*)blockStart;
int penalty = THRESHOLD_PENALTY;
size_t pos = 0;
assert(blockSize == (128 << 10));
assert(workspace != NULL);
assert((size_t)workspace % ZSTD_ALIGNOF(FPStats) == 0);
ZSTD_STATIC_ASSERT(ZSTD_SLIPBLOCK_WORKSPACESIZE >= sizeof(FPStats));
assert(wkspSize >= sizeof(FPStats)); (void)wkspSize;
initStats(fpstats);
record_f(&fpstats->pastEvents, p, CHUNKSIZE);
for (pos = CHUNKSIZE; pos <= blockSize - CHUNKSIZE; pos += CHUNKSIZE) {
record_f(&fpstats->newEvents, p + pos, CHUNKSIZE);
if (compareFingerprints(&fpstats->pastEvents, &fpstats->newEvents, penalty, hashParams[level])) {
return pos;
} else {
mergeEvents(&fpstats->pastEvents, &fpstats->newEvents);
if (penalty > 0) penalty--;
}
}
assert(pos == blockSize);
return blockSize;
(void)flushEvents; (void)removeEvents;
}
/* ZSTD_splitBlock_fromBorders(): very fast strategy :
* compare fingerprint from beginning and end of the block,
* derive from their difference if it's preferable to split in the middle,
* repeat the process a second time, for finer grained decision.
* 3 times did not brought improvements, so I stopped at 2.
* Benefits are good enough for a cheap heuristic.
* More accurate splitting saves more, but speed impact is also more perceptible.
* For better accuracy, use more elaborate variant *_byChunks.
*/
static size_t ZSTD_splitBlock_fromBorders(const void* blockStart, size_t blockSize,
void* workspace, size_t wkspSize)
{
#define SEGMENT_SIZE 512
FPStats* const fpstats = (FPStats*)workspace;
Fingerprint* middleEvents = (Fingerprint*)(void*)((char*)workspace + 512 * sizeof(unsigned));
assert(blockSize == (128 << 10));
assert(workspace != NULL);
assert((size_t)workspace % ZSTD_ALIGNOF(FPStats) == 0);
ZSTD_STATIC_ASSERT(ZSTD_SLIPBLOCK_WORKSPACESIZE >= sizeof(FPStats));
assert(wkspSize >= sizeof(FPStats)); (void)wkspSize;
initStats(fpstats);
HIST_add(fpstats->pastEvents.events, blockStart, SEGMENT_SIZE);
HIST_add(fpstats->newEvents.events, (const char*)blockStart + blockSize - SEGMENT_SIZE, SEGMENT_SIZE);
fpstats->pastEvents.nbEvents = fpstats->newEvents.nbEvents = SEGMENT_SIZE;
if (!compareFingerprints(&fpstats->pastEvents, &fpstats->newEvents, 0, 8))
return blockSize;
HIST_add(middleEvents->events, (const char*)blockStart + blockSize/2 - SEGMENT_SIZE/2, SEGMENT_SIZE);
middleEvents->nbEvents = SEGMENT_SIZE;
{ U64 const distFromBegin = fpDistance(&fpstats->pastEvents, middleEvents, 8);
U64 const distFromEnd = fpDistance(&fpstats->newEvents, middleEvents, 8);
U64 const minDistance = SEGMENT_SIZE * SEGMENT_SIZE / 3;
if (abs64((S64)distFromBegin - (S64)distFromEnd) < minDistance)
return 64 KB;
return (distFromBegin > distFromEnd) ? 32 KB : 96 KB;
}
}
size_t ZSTD_splitBlock(const void* blockStart, size_t blockSize,
int level,
void* workspace, size_t wkspSize)
{
DEBUGLOG(6, "ZSTD_splitBlock (level=%i)", level);
assert(0<=level && level<=4);
if (level == 0)
return ZSTD_splitBlock_fromBorders(blockStart, blockSize, workspace, wkspSize);
/* level >= 1*/
return ZSTD_splitBlock_byChunks(blockStart, blockSize, level-1, workspace, wkspSize);
}
Utilities/cmzstd/lib/compress/zstd_preSplit.h
+33
View File
@@ -0,0 +1,33 @@
/*
* Copyright (c) Meta Platforms, Inc. and affiliates.
* 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.
*/
#ifndef ZSTD_PRESPLIT_H
#define ZSTD_PRESPLIT_H
#include <stddef.h> /* size_t */
#define ZSTD_SLIPBLOCK_WORKSPACESIZE 8208
/* ZSTD_splitBlock():
* @level must be a value between 0 and 4.
* higher levels spend more energy to detect block boundaries.
* @workspace must be aligned for size_t.
* @wkspSize must be at least >= ZSTD_SLIPBLOCK_WORKSPACESIZE
* note:
* For the time being, this function only accepts full 128 KB blocks.
* Therefore, @blockSize must be == 128 KB.
* While this could be extended to smaller sizes in the future,
* it is not yet clear if this would be useful. TBD.
*/
size_t ZSTD_splitBlock(const void* blockStart, size_t blockSize,
int level,
void* workspace, size_t wkspSize);
#endif /* ZSTD_PRESPLIT_H */
Utilities/cmzstd/lib/compress/zstdmt_compress.c
+223 -167
View File
@@ -15,17 +15,13 @@
#endif
/* ====== Constants ====== */
#define ZSTDMT_OVERLAPLOG_DEFAULT 0
/* ====== Dependencies ====== */
#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
#include "../common/zstd_deps.h" /* ZSTD_memcpy, ZSTD_memset, INT_MAX, UINT_MAX */
#include "../common/mem.h" /* MEM_STATIC */
#include "../common/pool.h" /* threadpool */
#include "../common/threading.h" /* mutex */
#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */
#include "zstd_compress_internal.h" /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */
#include "zstd_ldm.h"
#include "zstdmt_compress.h"
@@ -44,12 +40,13 @@
# include <unistd.h>
# include <sys/times.h>
# define DEBUG_PRINTHEX(l,p,n) { \
unsigned debug_u; \
for (debug_u=0; debug_u<(n); debug_u++) \
RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \
RAWLOG(l, " \n"); \
}
# define DEBUG_PRINTHEX(l,p,n) \
do { \
unsigned debug_u; \
for (debug_u=0; debug_u<(n); debug_u++) \
RAWLOG(l, "%02X ", ((const unsigned char*)(p))[debug_u]); \
RAWLOG(l, " \n"); \
} while (0)
static unsigned long long GetCurrentClockTimeMicroseconds(void)
{
@@ -61,25 +58,28 @@ static unsigned long long GetCurrentClockTimeMicroseconds(void)
} }
#define MUTEX_WAIT_TIME_DLEVEL 6
#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) { \
if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) { \
unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \
ZSTD_pthread_mutex_lock(mutex); \
{ unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \
unsigned long long const elapsedTime = (afterTime-beforeTime); \
if (elapsedTime > 1000) { /* or whatever threshold you like; I'm using 1 millisecond here */ \
DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, "Thread took %llu microseconds to acquire mutex %s \n", \
elapsedTime, #mutex); \
} } \
} else { \
ZSTD_pthread_mutex_lock(mutex); \
} \
}
#define ZSTD_PTHREAD_MUTEX_LOCK(mutex) \
do { \
if (DEBUGLEVEL >= MUTEX_WAIT_TIME_DLEVEL) { \
unsigned long long const beforeTime = GetCurrentClockTimeMicroseconds(); \
ZSTD_pthread_mutex_lock(mutex); \
{ unsigned long long const afterTime = GetCurrentClockTimeMicroseconds(); \
unsigned long long const elapsedTime = (afterTime-beforeTime); \
if (elapsedTime > 1000) { \
/* or whatever threshold you like; I'm using 1 millisecond here */ \
DEBUGLOG(MUTEX_WAIT_TIME_DLEVEL, \
"Thread took %llu microseconds to acquire mutex %s \n", \
elapsedTime, #mutex); \
} } \
} else { \
ZSTD_pthread_mutex_lock(mutex); \
} \
} while (0)
#else
# define ZSTD_PTHREAD_MUTEX_LOCK(m) ZSTD_pthread_mutex_lock(m)
# define DEBUG_PRINTHEX(l,p,n) {}
# define DEBUG_PRINTHEX(l,p,n) do { } while (0)
#endif
@@ -90,9 +90,9 @@ static unsigned long long GetCurrentClockTimeMicroseconds(void)
typedef struct buffer_s {
void* start;
size_t capacity;
} buffer_t;
} Buffer;
static const buffer_t g_nullBuffer = { NULL, 0 };
static const Buffer g_nullBuffer = { NULL, 0 };
typedef struct ZSTDMT_bufferPool_s {
ZSTD_pthread_mutex_t poolMutex;
@@ -100,18 +100,39 @@ typedef struct ZSTDMT_bufferPool_s {
unsigned totalBuffers;
unsigned nbBuffers;
ZSTD_customMem cMem;
buffer_t bTable[1]; /* variable size */
Buffer* buffers;
} ZSTDMT_bufferPool;
static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)
{
DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);
if (!bufPool) return; /* compatibility with free on NULL */
if (bufPool->buffers) {
unsigned u;
for (u=0; u<bufPool->totalBuffers; u++) {
DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->buffers[u].start);
ZSTD_customFree(bufPool->buffers[u].start, bufPool->cMem);
}
ZSTD_customFree(bufPool->buffers, bufPool->cMem);
}
ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);
ZSTD_customFree(bufPool, bufPool->cMem);
}
static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_customMem cMem)
{
ZSTDMT_bufferPool* const bufPool = (ZSTDMT_bufferPool*)ZSTD_customCalloc(
sizeof(ZSTDMT_bufferPool) + (maxNbBuffers-1) * sizeof(buffer_t), cMem);
ZSTDMT_bufferPool* const bufPool =
(ZSTDMT_bufferPool*)ZSTD_customCalloc(sizeof(ZSTDMT_bufferPool), cMem);
if (bufPool==NULL) return NULL;
if (ZSTD_pthread_mutex_init(&bufPool->poolMutex, NULL)) {
ZSTD_customFree(bufPool, cMem);
return NULL;
}
bufPool->buffers = (Buffer*)ZSTD_customCalloc(maxNbBuffers * sizeof(Buffer), cMem);
if (bufPool->buffers==NULL) {
ZSTDMT_freeBufferPool(bufPool);
return NULL;
}
bufPool->bufferSize = 64 KB;
bufPool->totalBuffers = maxNbBuffers;
bufPool->nbBuffers = 0;
@@ -119,32 +140,19 @@ static ZSTDMT_bufferPool* ZSTDMT_createBufferPool(unsigned maxNbBuffers, ZSTD_cu
return bufPool;
}
static void ZSTDMT_freeBufferPool(ZSTDMT_bufferPool* bufPool)
{
unsigned u;
DEBUGLOG(3, "ZSTDMT_freeBufferPool (address:%08X)", (U32)(size_t)bufPool);
if (!bufPool) return; /* compatibility with free on NULL */
for (u=0; u<bufPool->totalBuffers; u++) {
DEBUGLOG(4, "free buffer %2u (address:%08X)", u, (U32)(size_t)bufPool->bTable[u].start);
ZSTD_customFree(bufPool->bTable[u].start, bufPool->cMem);
}
ZSTD_pthread_mutex_destroy(&bufPool->poolMutex);
ZSTD_customFree(bufPool, bufPool->cMem);
}
/* only works at initialization, not during compression */
static size_t ZSTDMT_sizeof_bufferPool(ZSTDMT_bufferPool* bufPool)
{
size_t const poolSize = sizeof(*bufPool)
+ (bufPool->totalBuffers - 1) * sizeof(buffer_t);
size_t const poolSize = sizeof(*bufPool);
size_t const arraySize = bufPool->totalBuffers * sizeof(Buffer);
unsigned u;
size_t totalBufferSize = 0;
ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
for (u=0; u<bufPool->totalBuffers; u++)
totalBufferSize += bufPool->bTable[u].capacity;
totalBufferSize += bufPool->buffers[u].capacity;
ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
return poolSize + totalBufferSize;
return poolSize + arraySize + totalBufferSize;
}
/* ZSTDMT_setBufferSize() :
@@ -181,15 +189,15 @@ static ZSTDMT_bufferPool* ZSTDMT_expandBufferPool(ZSTDMT_bufferPool* srcBufPool,
* assumption : bufPool must be valid
* @return : a buffer, with start pointer and size
* note: allocation may fail, in this case, start==NULL and size==0 */
static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
static Buffer ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
{
size_t const bSize = bufPool->bufferSize;
DEBUGLOG(5, "ZSTDMT_getBuffer: bSize = %u", (U32)bufPool->bufferSize);
ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
if (bufPool->nbBuffers) { /* try to use an existing buffer */
buffer_t const buf = bufPool->bTable[--(bufPool->nbBuffers)];
Buffer const buf = bufPool->buffers[--(bufPool->nbBuffers)];
size_t const availBufferSize = buf.capacity;
bufPool->bTable[bufPool->nbBuffers] = g_nullBuffer;
bufPool->buffers[bufPool->nbBuffers] = g_nullBuffer;
if ((availBufferSize >= bSize) & ((availBufferSize>>3) <= bSize)) {
/* large enough, but not too much */
DEBUGLOG(5, "ZSTDMT_getBuffer: provide buffer %u of size %u",
@@ -204,7 +212,7 @@ static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
/* create new buffer */
DEBUGLOG(5, "ZSTDMT_getBuffer: create a new buffer");
{ buffer_t buffer;
{ Buffer buffer;
void* const start = ZSTD_customMalloc(bSize, bufPool->cMem);
buffer.start = start; /* note : start can be NULL if malloc fails ! */
buffer.capacity = (start==NULL) ? 0 : bSize;
@@ -223,12 +231,12 @@ static buffer_t ZSTDMT_getBuffer(ZSTDMT_bufferPool* bufPool)
* @return : a buffer that is at least the buffer pool buffer size.
* If a reallocation happens, the data in the input buffer is copied.
*/
static buffer_t ZSTDMT_resizeBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buffer)
static Buffer ZSTDMT_resizeBuffer(ZSTDMT_bufferPool* bufPool, Buffer buffer)
{
size_t const bSize = bufPool->bufferSize;
if (buffer.capacity < bSize) {
void* const start = ZSTD_customMalloc(bSize, bufPool->cMem);
buffer_t newBuffer;
Buffer newBuffer;
newBuffer.start = start;
newBuffer.capacity = start == NULL ? 0 : bSize;
if (start != NULL) {
@@ -244,20 +252,20 @@ static buffer_t ZSTDMT_resizeBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buffer)
#endif
/* store buffer for later re-use, up to pool capacity */
static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, buffer_t buf)
static void ZSTDMT_releaseBuffer(ZSTDMT_bufferPool* bufPool, Buffer buf)
{
DEBUGLOG(5, "ZSTDMT_releaseBuffer");
if (buf.start == NULL) return; /* compatible with release on NULL */
ZSTD_pthread_mutex_lock(&bufPool->poolMutex);
if (bufPool->nbBuffers < bufPool->totalBuffers) {
bufPool->bTable[bufPool->nbBuffers++] = buf; /* stored for later use */
bufPool->buffers[bufPool->nbBuffers++] = buf; /* stored for later use */
DEBUGLOG(5, "ZSTDMT_releaseBuffer: stored buffer of size %u in slot %u",
(U32)buf.capacity, (U32)(bufPool->nbBuffers-1));
ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
return;
}
ZSTD_pthread_mutex_unlock(&bufPool->poolMutex);
/* Reached bufferPool capacity (should not happen) */
/* Reached bufferPool capacity (note: should not happen) */
DEBUGLOG(5, "ZSTDMT_releaseBuffer: pool capacity reached => freeing ");
ZSTD_customFree(buf.start, bufPool->cMem);
}
@@ -282,23 +290,23 @@ static size_t ZSTDMT_sizeof_seqPool(ZSTDMT_seqPool* seqPool)
return ZSTDMT_sizeof_bufferPool(seqPool);
}
static rawSeqStore_t bufferToSeq(buffer_t buffer)
static RawSeqStore_t bufferToSeq(Buffer buffer)
{
rawSeqStore_t seq = kNullRawSeqStore;
RawSeqStore_t seq = kNullRawSeqStore;
seq.seq = (rawSeq*)buffer.start;
seq.capacity = buffer.capacity / sizeof(rawSeq);
return seq;
}
static buffer_t seqToBuffer(rawSeqStore_t seq)
static Buffer seqToBuffer(RawSeqStore_t seq)
{
buffer_t buffer;
Buffer buffer;
buffer.start = seq.seq;
buffer.capacity = seq.capacity * sizeof(rawSeq);
return buffer;
}
static rawSeqStore_t ZSTDMT_getSeq(ZSTDMT_seqPool* seqPool)
static RawSeqStore_t ZSTDMT_getSeq(ZSTDMT_seqPool* seqPool)
{
if (seqPool->bufferSize == 0) {
return kNullRawSeqStore;
@@ -307,13 +315,13 @@ static rawSeqStore_t ZSTDMT_getSeq(ZSTDMT_seqPool* seqPool)
}
#if ZSTD_RESIZE_SEQPOOL
static rawSeqStore_t ZSTDMT_resizeSeq(ZSTDMT_seqPool* seqPool, rawSeqStore_t seq)
static RawSeqStore_t ZSTDMT_resizeSeq(ZSTDMT_seqPool* seqPool, RawSeqStore_t seq)
{
return bufferToSeq(ZSTDMT_resizeBuffer(seqPool, seqToBuffer(seq)));
}
#endif
static void ZSTDMT_releaseSeq(ZSTDMT_seqPool* seqPool, rawSeqStore_t seq)
static void ZSTDMT_releaseSeq(ZSTDMT_seqPool* seqPool, RawSeqStore_t seq)
{
ZSTDMT_releaseBuffer(seqPool, seqToBuffer(seq));
}
@@ -350,16 +358,20 @@ typedef struct {
int totalCCtx;
int availCCtx;
ZSTD_customMem cMem;
ZSTD_CCtx* cctx[1]; /* variable size */
ZSTD_CCtx** cctxs;
} ZSTDMT_CCtxPool;
/* note : all CCtx borrowed from the pool should be released back to the pool _before_ freeing the pool */
/* note : all CCtx borrowed from the pool must be reverted back to the pool _before_ freeing the pool */
static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)
{
int cid;
for (cid=0; cid<pool->totalCCtx; cid++)
ZSTD_freeCCtx(pool->cctx[cid]); /* note : compatible with free on NULL */
if (!pool) return;
ZSTD_pthread_mutex_destroy(&pool->poolMutex);
if (pool->cctxs) {
int cid;
for (cid=0; cid<pool->totalCCtx; cid++)
ZSTD_freeCCtx(pool->cctxs[cid]); /* free compatible with NULL */
ZSTD_customFree(pool->cctxs, pool->cMem);
}
ZSTD_customFree(pool, pool->cMem);
}
@@ -368,19 +380,24 @@ static void ZSTDMT_freeCCtxPool(ZSTDMT_CCtxPool* pool)
static ZSTDMT_CCtxPool* ZSTDMT_createCCtxPool(int nbWorkers,
ZSTD_customMem cMem)
{
ZSTDMT_CCtxPool* const cctxPool = (ZSTDMT_CCtxPool*) ZSTD_customCalloc(
sizeof(ZSTDMT_CCtxPool) + (nbWorkers-1)*sizeof(ZSTD_CCtx*), cMem);
ZSTDMT_CCtxPool* const cctxPool =
(ZSTDMT_CCtxPool*) ZSTD_customCalloc(sizeof(ZSTDMT_CCtxPool), cMem);
assert(nbWorkers > 0);
if (!cctxPool) return NULL;
if (ZSTD_pthread_mutex_init(&cctxPool->poolMutex, NULL)) {
ZSTD_customFree(cctxPool, cMem);
return NULL;
}
cctxPool->cMem = cMem;
cctxPool->totalCCtx = nbWorkers;
cctxPool->cctxs = (ZSTD_CCtx**)ZSTD_customCalloc(nbWorkers * sizeof(ZSTD_CCtx*), cMem);
if (!cctxPool->cctxs) {
ZSTDMT_freeCCtxPool(cctxPool);
return NULL;
}
cctxPool->cMem = cMem;
cctxPool->cctxs[0] = ZSTD_createCCtx_advanced(cMem);
if (!cctxPool->cctxs[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; }
cctxPool->availCCtx = 1; /* at least one cctx for single-thread mode */
cctxPool->cctx[0] = ZSTD_createCCtx_advanced(cMem);
if (!cctxPool->cctx[0]) { ZSTDMT_freeCCtxPool(cctxPool); return NULL; }
DEBUGLOG(3, "cctxPool created, with %u workers", nbWorkers);
return cctxPool;
}
@@ -402,16 +419,16 @@ static size_t ZSTDMT_sizeof_CCtxPool(ZSTDMT_CCtxPool* cctxPool)
{
ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
{ unsigned const nbWorkers = cctxPool->totalCCtx;
size_t const poolSize = sizeof(*cctxPool)
+ (nbWorkers-1) * sizeof(ZSTD_CCtx*);
unsigned u;
size_t const poolSize = sizeof(*cctxPool);
size_t const arraySize = cctxPool->totalCCtx * sizeof(ZSTD_CCtx*);
size_t totalCCtxSize = 0;
unsigned u;
for (u=0; u<nbWorkers; u++) {
totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctx[u]);
totalCCtxSize += ZSTD_sizeof_CCtx(cctxPool->cctxs[u]);
}
ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
assert(nbWorkers > 0);
return poolSize + totalCCtxSize;
return poolSize + arraySize + totalCCtxSize;
}
}
@@ -421,7 +438,7 @@ static ZSTD_CCtx* ZSTDMT_getCCtx(ZSTDMT_CCtxPool* cctxPool)
ZSTD_pthread_mutex_lock(&cctxPool->poolMutex);
if (cctxPool->availCCtx) {
cctxPool->availCCtx--;
{ ZSTD_CCtx* const cctx = cctxPool->cctx[cctxPool->availCCtx];
{ ZSTD_CCtx* const cctx = cctxPool->cctxs[cctxPool->availCCtx];
ZSTD_pthread_mutex_unlock(&cctxPool->poolMutex);
return cctx;
} }
@@ -435,7 +452,7 @@ static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)
if (cctx==NULL) return; /* compatibility with release on NULL */
ZSTD_pthread_mutex_lock(&pool->poolMutex);
if (pool->availCCtx < pool->totalCCtx)
pool->cctx[pool->availCCtx++] = cctx;
pool->cctxs[pool->availCCtx++] = cctx;
else {
/* pool overflow : should not happen, since totalCCtx==nbWorkers */
DEBUGLOG(4, "CCtx pool overflow : free cctx");
@@ -449,7 +466,7 @@ static void ZSTDMT_releaseCCtx(ZSTDMT_CCtxPool* pool, ZSTD_CCtx* cctx)
typedef struct {
void const* start;
size_t size;
} range_t;
} Range;
typedef struct {
/* All variables in the struct are protected by mutex. */
@@ -465,10 +482,10 @@ typedef struct {
ZSTD_pthread_mutex_t ldmWindowMutex;
ZSTD_pthread_cond_t ldmWindowCond; /* Signaled when ldmWindow is updated */
ZSTD_window_t ldmWindow; /* A thread-safe copy of ldmState.window */
} serialState_t;
} SerialState;
static int
ZSTDMT_serialState_reset(serialState_t* serialState,
ZSTDMT_serialState_reset(SerialState* serialState,
ZSTDMT_seqPool* seqPool,
ZSTD_CCtx_params params,
size_t jobSize,
@@ -538,7 +555,7 @@ ZSTDMT_serialState_reset(serialState_t* serialState,
return 0;
}
static int ZSTDMT_serialState_init(serialState_t* serialState)
static int ZSTDMT_serialState_init(SerialState* serialState)
{
int initError = 0;
ZSTD_memset(serialState, 0, sizeof(*serialState));
@@ -549,7 +566,7 @@ static int ZSTDMT_serialState_init(serialState_t* serialState)
return initError;
}
static void ZSTDMT_serialState_free(serialState_t* serialState)
static void ZSTDMT_serialState_free(SerialState* serialState)
{
ZSTD_customMem cMem = serialState->params.customMem;
ZSTD_pthread_mutex_destroy(&serialState->mutex);
@@ -560,9 +577,10 @@ static void ZSTDMT_serialState_free(serialState_t* serialState)
ZSTD_customFree(serialState->ldmState.bucketOffsets, cMem);
}
static void ZSTDMT_serialState_update(serialState_t* serialState,
ZSTD_CCtx* jobCCtx, rawSeqStore_t seqStore,
range_t src, unsigned jobID)
static void
ZSTDMT_serialState_genSequences(SerialState* serialState,
RawSeqStore_t* seqStore,
Range src, unsigned jobID)
{
/* Wait for our turn */
ZSTD_PTHREAD_MUTEX_LOCK(&serialState->mutex);
@@ -575,12 +593,13 @@ static void ZSTDMT_serialState_update(serialState_t* serialState,
/* It is now our turn, do any processing necessary */
if (serialState->params.ldmParams.enableLdm == ZSTD_ps_enable) {
size_t error;
assert(seqStore.seq != NULL && seqStore.pos == 0 &&
seqStore.size == 0 && seqStore.capacity > 0);
DEBUGLOG(6, "ZSTDMT_serialState_genSequences: LDM update");
assert(seqStore->seq != NULL && seqStore->pos == 0 &&
seqStore->size == 0 && seqStore->capacity > 0);
assert(src.size <= serialState->params.jobSize);
ZSTD_window_update(&serialState->ldmState.window, src.start, src.size, /* forceNonContiguous */ 0);
error = ZSTD_ldm_generateSequences(
&serialState->ldmState, &seqStore,
&serialState->ldmState, seqStore,
&serialState->params.ldmParams, src.start, src.size);
/* We provide a large enough buffer to never fail. */
assert(!ZSTD_isError(error)); (void)error;
@@ -599,17 +618,22 @@ static void ZSTDMT_serialState_update(serialState_t* serialState,
serialState->nextJobID++;
ZSTD_pthread_cond_broadcast(&serialState->cond);
ZSTD_pthread_mutex_unlock(&serialState->mutex);
}
if (seqStore.size > 0) {
size_t const err = ZSTD_referenceExternalSequences(
jobCCtx, seqStore.seq, seqStore.size);
assert(serialState->params.ldmParams.enableLdm == ZSTD_ps_enable);
assert(!ZSTD_isError(err));
(void)err;
static void
ZSTDMT_serialState_applySequences(const SerialState* serialState, /* just for an assert() check */
ZSTD_CCtx* jobCCtx,
const RawSeqStore_t* seqStore)
{
if (seqStore->size > 0) {
DEBUGLOG(5, "ZSTDMT_serialState_applySequences: uploading %u external sequences", (unsigned)seqStore->size);
assert(serialState->params.ldmParams.enableLdm == ZSTD_ps_enable); (void)serialState;
assert(jobCCtx);
ZSTD_referenceExternalSequences(jobCCtx, seqStore->seq, seqStore->size);
}
}
static void ZSTDMT_serialState_ensureFinished(serialState_t* serialState,
static void ZSTDMT_serialState_ensureFinished(SerialState* serialState,
unsigned jobID, size_t cSize)
{
ZSTD_PTHREAD_MUTEX_LOCK(&serialState->mutex);
@@ -633,36 +657,37 @@ static void ZSTDMT_serialState_ensureFinished(serialState_t* serialState,
/* ===== Worker thread ===== */
/* ------------------------------------------ */
static const range_t kNullRange = { NULL, 0 };
static const Range kNullRange = { NULL, 0 };
typedef struct {
size_t consumed; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx */
size_t cSize; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx, then set0 by mtctx */
ZSTD_pthread_mutex_t job_mutex; /* Thread-safe - used by mtctx and worker */
ZSTD_pthread_cond_t job_cond; /* Thread-safe - used by mtctx and worker */
ZSTDMT_CCtxPool* cctxPool; /* Thread-safe - used by mtctx and (all) workers */
ZSTDMT_bufferPool* bufPool; /* Thread-safe - used by mtctx and (all) workers */
ZSTDMT_seqPool* seqPool; /* Thread-safe - used by mtctx and (all) workers */
serialState_t* serial; /* Thread-safe - used by mtctx and (all) workers */
buffer_t dstBuff; /* set by worker (or mtctx), then read by worker & mtctx, then modified by mtctx => no barrier */
range_t prefix; /* set by mtctx, then read by worker & mtctx => no barrier */
range_t src; /* set by mtctx, then read by worker & mtctx => no barrier */
unsigned jobID; /* set by mtctx, then read by worker => no barrier */
unsigned firstJob; /* set by mtctx, then read by worker => no barrier */
unsigned lastJob; /* set by mtctx, then read by worker => no barrier */
ZSTD_CCtx_params params; /* set by mtctx, then read by worker => no barrier */
const ZSTD_CDict* cdict; /* set by mtctx, then read by worker => no barrier */
unsigned long long fullFrameSize; /* set by mtctx, then read by worker => no barrier */
size_t dstFlushed; /* used only by mtctx */
unsigned frameChecksumNeeded; /* used only by mtctx */
size_t consumed; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx */
size_t cSize; /* SHARED - set0 by mtctx, then modified by worker AND read by mtctx, then set0 by mtctx */
ZSTD_pthread_mutex_t job_mutex; /* Thread-safe - used by mtctx and worker */
ZSTD_pthread_cond_t job_cond; /* Thread-safe - used by mtctx and worker */
ZSTDMT_CCtxPool* cctxPool; /* Thread-safe - used by mtctx and (all) workers */
ZSTDMT_bufferPool* bufPool; /* Thread-safe - used by mtctx and (all) workers */
ZSTDMT_seqPool* seqPool; /* Thread-safe - used by mtctx and (all) workers */
SerialState* serial; /* Thread-safe - used by mtctx and (all) workers */
Buffer dstBuff; /* set by worker (or mtctx), then read by worker & mtctx, then modified by mtctx => no barrier */
Range prefix; /* set by mtctx, then read by worker & mtctx => no barrier */
Range src; /* set by mtctx, then read by worker & mtctx => no barrier */
unsigned jobID; /* set by mtctx, then read by worker => no barrier */
unsigned firstJob; /* set by mtctx, then read by worker => no barrier */
unsigned lastJob; /* set by mtctx, then read by worker => no barrier */
ZSTD_CCtx_params params; /* set by mtctx, then read by worker => no barrier */
const ZSTD_CDict* cdict; /* set by mtctx, then read by worker => no barrier */
unsigned long long fullFrameSize; /* set by mtctx, then read by worker => no barrier */
size_t dstFlushed; /* used only by mtctx */
unsigned frameChecksumNeeded; /* used only by mtctx */
} ZSTDMT_jobDescription;
#define JOB_ERROR(e) { \
ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex); \
job->cSize = e; \
ZSTD_pthread_mutex_unlock(&job->job_mutex); \
goto _endJob; \
}
#define JOB_ERROR(e) \
do { \
ZSTD_PTHREAD_MUTEX_LOCK(&job->job_mutex); \
job->cSize = e; \
ZSTD_pthread_mutex_unlock(&job->job_mutex); \
goto _endJob; \
} while (0)
/* ZSTDMT_compressionJob() is a POOL_function type */
static void ZSTDMT_compressionJob(void* jobDescription)
@@ -670,10 +695,11 @@ static void ZSTDMT_compressionJob(void* jobDescription)
ZSTDMT_jobDescription* const job = (ZSTDMT_jobDescription*)jobDescription;
ZSTD_CCtx_params jobParams = job->params; /* do not modify job->params ! copy it, modify the copy */
ZSTD_CCtx* const cctx = ZSTDMT_getCCtx(job->cctxPool);
rawSeqStore_t rawSeqStore = ZSTDMT_getSeq(job->seqPool);
buffer_t dstBuff = job->dstBuff;
RawSeqStore_t rawSeqStore = ZSTDMT_getSeq(job->seqPool);
Buffer dstBuff = job->dstBuff;
size_t lastCBlockSize = 0;
DEBUGLOG(5, "ZSTDMT_compressionJob: job %u", job->jobID);
/* resources */
if (cctx==NULL) JOB_ERROR(ERROR(memory_allocation));
if (dstBuff.start == NULL) { /* streaming job : doesn't provide a dstBuffer */
@@ -695,11 +721,15 @@ static void ZSTDMT_compressionJob(void* jobDescription)
/* init */
/* Perform serial step as early as possible */
ZSTDMT_serialState_genSequences(job->serial, &rawSeqStore, job->src, job->jobID);
if (job->cdict) {
size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast, job->cdict, &jobParams, job->fullFrameSize);
assert(job->firstJob); /* only allowed for first job */
if (ZSTD_isError(initError)) JOB_ERROR(initError);
} else { /* srcStart points at reloaded section */
} else {
U64 const pledgedSrcSize = job->firstJob ? job->fullFrameSize : job->src.size;
{ size_t const forceWindowError = ZSTD_CCtxParams_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob);
if (ZSTD_isError(forceWindowError)) JOB_ERROR(forceWindowError);
@@ -708,16 +738,17 @@ static void ZSTDMT_compressionJob(void* jobDescription)
size_t const err = ZSTD_CCtxParams_setParameter(&jobParams, ZSTD_c_deterministicRefPrefix, 0);
if (ZSTD_isError(err)) JOB_ERROR(err);
}
DEBUGLOG(6, "ZSTDMT_compressionJob: job %u: loading prefix of size %zu", job->jobID, job->prefix.size);
{ size_t const initError = ZSTD_compressBegin_advanced_internal(cctx,
job->prefix.start, job->prefix.size, ZSTD_dct_rawContent, /* load dictionary in "content-only" mode (no header analysis) */
job->prefix.start, job->prefix.size, ZSTD_dct_rawContent,
ZSTD_dtlm_fast,
NULL, /*cdict*/
&jobParams, pledgedSrcSize);
if (ZSTD_isError(initError)) JOB_ERROR(initError);
} }
/* Perform serial step as early as possible, but after CCtx initialization */
ZSTDMT_serialState_update(job->serial, cctx, rawSeqStore, job->src, job->jobID);
/* External Sequences can only be applied after CCtx initialization */
ZSTDMT_serialState_applySequences(job->serial, cctx, &rawSeqStore);
if (!job->firstJob) { /* flush and overwrite frame header when it's not first job */
size_t const hSize = ZSTD_compressContinue_public(cctx, dstBuff.start, dstBuff.capacity, job->src.start, 0);
@@ -726,7 +757,7 @@ static void ZSTDMT_compressionJob(void* jobDescription)
ZSTD_invalidateRepCodes(cctx);
}
/* compress */
/* compress the entire job by smaller chunks, for better granularity */
{ size_t const chunkSize = 4*ZSTD_BLOCKSIZE_MAX;
int const nbChunks = (int)((job->src.size + (chunkSize-1)) / chunkSize);
const BYTE* ip = (const BYTE*) job->src.start;
@@ -794,10 +825,10 @@ _endJob:
/* ------------------------------------------ */
typedef struct {
range_t prefix; /* read-only non-owned prefix buffer */
buffer_t buffer;
Range prefix; /* read-only non-owned prefix buffer */
Buffer buffer;
size_t filled;
} inBuff_t;
} InBuff_t;
typedef struct {
BYTE* buffer; /* The round input buffer. All jobs get references
@@ -811,9 +842,9 @@ typedef struct {
* the inBuff is sent to the worker thread.
* pos <= capacity.
*/
} roundBuff_t;
} RoundBuff_t;
static const roundBuff_t kNullRoundBuff = {NULL, 0, 0};
static const RoundBuff_t kNullRoundBuff = {NULL, 0, 0};
#define RSYNC_LENGTH 32
/* Don't create chunks smaller than the zstd block size.
@@ -830,7 +861,7 @@ typedef struct {
U64 hash;
U64 hitMask;
U64 primePower;
} rsyncState_t;
} RSyncState_t;
struct ZSTDMT_CCtx_s {
POOL_ctx* factory;
@@ -842,10 +873,10 @@ struct ZSTDMT_CCtx_s {
size_t targetSectionSize;
size_t targetPrefixSize;
int jobReady; /* 1 => one job is already prepared, but pool has shortage of workers. Don't create a new job. */
inBuff_t inBuff;
roundBuff_t roundBuff;
serialState_t serial;
rsyncState_t rsync;
InBuff_t inBuff;
RoundBuff_t roundBuff;
SerialState serial;
RSyncState_t rsync;
unsigned jobIDMask;
unsigned doneJobID;
unsigned nextJobID;
@@ -1091,7 +1122,7 @@ ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx)
{ unsigned jobNb;
unsigned lastJobNb = mtctx->nextJobID + mtctx->jobReady; assert(mtctx->jobReady <= 1);
DEBUGLOG(6, "ZSTDMT_getFrameProgression: jobs: from %u to <%u (jobReady:%u)",
mtctx->doneJobID, lastJobNb, mtctx->jobReady)
mtctx->doneJobID, lastJobNb, mtctx->jobReady);
for (jobNb = mtctx->doneJobID ; jobNb < lastJobNb ; jobNb++) {
unsigned const wJobID = jobNb & mtctx->jobIDMask;
ZSTDMT_jobDescription* jobPtr = &mtctx->jobs[wJobID];
@@ -1230,13 +1261,11 @@ size_t ZSTDMT_initCStream_internal(
/* init */
if (params.nbWorkers != mtctx->params.nbWorkers)
FORWARD_IF_ERROR( ZSTDMT_resize(mtctx, params.nbWorkers) , "");
FORWARD_IF_ERROR( ZSTDMT_resize(mtctx, (unsigned)params.nbWorkers) , "");
if (params.jobSize != 0 && params.jobSize < ZSTDMT_JOBSIZE_MIN) params.jobSize = ZSTDMT_JOBSIZE_MIN;
if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = (size_t)ZSTDMT_JOBSIZE_MAX;
DEBUGLOG(4, "ZSTDMT_initCStream_internal: %u workers", params.nbWorkers);
if (mtctx->allJobsCompleted == 0) { /* previous compression not correctly finished */
ZSTDMT_waitForAllJobsCompleted(mtctx);
ZSTDMT_releaseAllJobResources(mtctx);
@@ -1245,15 +1274,14 @@ size_t ZSTDMT_initCStream_internal(
mtctx->params = params;
mtctx->frameContentSize = pledgedSrcSize;
ZSTD_freeCDict(mtctx->cdictLocal);
if (dict) {
ZSTD_freeCDict(mtctx->cdictLocal);
mtctx->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
ZSTD_dlm_byCopy, dictContentType, /* note : a loadPrefix becomes an internal CDict */
params.cParams, mtctx->cMem);
mtctx->cdict = mtctx->cdictLocal;
if (mtctx->cdictLocal == NULL) return ERROR(memory_allocation);
} else {
ZSTD_freeCDict(mtctx->cdictLocal);
mtctx->cdictLocal = NULL;
mtctx->cdict = cdict;
}
@@ -1319,9 +1347,32 @@ size_t ZSTDMT_initCStream_internal(
mtctx->allJobsCompleted = 0;
mtctx->consumed = 0;
mtctx->produced = 0;
/* update dictionary */
ZSTD_freeCDict(mtctx->cdictLocal);
mtctx->cdictLocal = NULL;
mtctx->cdict = NULL;
if (dict) {
if (dictContentType == ZSTD_dct_rawContent) {
mtctx->inBuff.prefix.start = (const BYTE*)dict;
mtctx->inBuff.prefix.size = dictSize;
} else {
/* note : a loadPrefix becomes an internal CDict */
mtctx->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
ZSTD_dlm_byRef, dictContentType,
params.cParams, mtctx->cMem);
mtctx->cdict = mtctx->cdictLocal;
if (mtctx->cdictLocal == NULL) return ERROR(memory_allocation);
}
} else {
mtctx->cdict = cdict;
}
if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, mtctx->targetSectionSize,
dict, dictSize, dictContentType))
return ERROR(memory_allocation);
return 0;
}
@@ -1388,7 +1439,7 @@ static size_t ZSTDMT_createCompressionJob(ZSTDMT_CCtx* mtctx, size_t srcSize, ZS
mtctx->roundBuff.pos += srcSize;
mtctx->inBuff.buffer = g_nullBuffer;
mtctx->inBuff.filled = 0;
/* Set the prefix */
/* Set the prefix for next job */
if (!endFrame) {
size_t const newPrefixSize = MIN(srcSize, mtctx->targetPrefixSize);
mtctx->inBuff.prefix.start = src + srcSize - newPrefixSize;
@@ -1525,12 +1576,17 @@ static size_t ZSTDMT_flushProduced(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, u
* If the data of the first job is broken up into two segments, we cover both
* sections.
*/
static range_t ZSTDMT_getInputDataInUse(ZSTDMT_CCtx* mtctx)
static Range ZSTDMT_getInputDataInUse(ZSTDMT_CCtx* mtctx)
{
unsigned const firstJobID = mtctx->doneJobID;
unsigned const lastJobID = mtctx->nextJobID;
unsigned jobID;
/* no need to check during first round */
size_t roundBuffCapacity = mtctx->roundBuff.capacity;
size_t nbJobs1stRoundMin = roundBuffCapacity / mtctx->targetSectionSize;
if (lastJobID < nbJobs1stRoundMin) return kNullRange;
for (jobID = firstJobID; jobID < lastJobID; ++jobID) {
unsigned const wJobID = jobID & mtctx->jobIDMask;
size_t consumed;
@@ -1540,7 +1596,7 @@ static range_t ZSTDMT_getInputDataInUse(ZSTDMT_CCtx* mtctx)
ZSTD_pthread_mutex_unlock(&mtctx->jobs[wJobID].job_mutex);
if (consumed < mtctx->jobs[wJobID].src.size) {
range_t range = mtctx->jobs[wJobID].prefix;
Range range = mtctx->jobs[wJobID].prefix;
if (range.size == 0) {
/* Empty prefix */
range = mtctx->jobs[wJobID].src;
@@ -1556,7 +1612,7 @@ static range_t ZSTDMT_getInputDataInUse(ZSTDMT_CCtx* mtctx)
/**
* Returns non-zero iff buffer and range overlap.
*/
static int ZSTDMT_isOverlapped(buffer_t buffer, range_t range)
static int ZSTDMT_isOverlapped(Buffer buffer, Range range)
{
BYTE const* const bufferStart = (BYTE const*)buffer.start;
BYTE const* const rangeStart = (BYTE const*)range.start;
@@ -1576,10 +1632,10 @@ static int ZSTDMT_isOverlapped(buffer_t buffer, range_t range)
}
}
static int ZSTDMT_doesOverlapWindow(buffer_t buffer, ZSTD_window_t window)
static int ZSTDMT_doesOverlapWindow(Buffer buffer, ZSTD_window_t window)
{
range_t extDict;
range_t prefix;
Range extDict;
Range prefix;
DEBUGLOG(5, "ZSTDMT_doesOverlapWindow");
extDict.start = window.dictBase + window.lowLimit;
@@ -1598,7 +1654,7 @@ static int ZSTDMT_doesOverlapWindow(buffer_t buffer, ZSTD_window_t window)
|| ZSTDMT_isOverlapped(buffer, prefix);
}
static void ZSTDMT_waitForLdmComplete(ZSTDMT_CCtx* mtctx, buffer_t buffer)
static void ZSTDMT_waitForLdmComplete(ZSTDMT_CCtx* mtctx, Buffer buffer)
{
if (mtctx->params.ldmParams.enableLdm == ZSTD_ps_enable) {
ZSTD_pthread_mutex_t* mutex = &mtctx->serial.ldmWindowMutex;
@@ -1623,16 +1679,16 @@ static void ZSTDMT_waitForLdmComplete(ZSTDMT_CCtx* mtctx, buffer_t buffer)
*/
static int ZSTDMT_tryGetInputRange(ZSTDMT_CCtx* mtctx)
{
range_t const inUse = ZSTDMT_getInputDataInUse(mtctx);
Range const inUse = ZSTDMT_getInputDataInUse(mtctx);
size_t const spaceLeft = mtctx->roundBuff.capacity - mtctx->roundBuff.pos;
size_t const target = mtctx->targetSectionSize;
buffer_t buffer;
size_t const spaceNeeded = mtctx->targetSectionSize;
Buffer buffer;
DEBUGLOG(5, "ZSTDMT_tryGetInputRange");
assert(mtctx->inBuff.buffer.start == NULL);
assert(mtctx->roundBuff.capacity >= target);
assert(mtctx->roundBuff.capacity >= spaceNeeded);
if (spaceLeft < target) {
if (spaceLeft < spaceNeeded) {
/* ZSTD_invalidateRepCodes() doesn't work for extDict variants.
* Simply copy the prefix to the beginning in that case.
*/
@@ -1651,7 +1707,7 @@ static int ZSTDMT_tryGetInputRange(ZSTDMT_CCtx* mtctx)
mtctx->roundBuff.pos = prefixSize;
}
buffer.start = mtctx->roundBuff.buffer + mtctx->roundBuff.pos;
buffer.capacity = target;
buffer.capacity = spaceNeeded;
if (ZSTDMT_isOverlapped(buffer, inUse)) {
DEBUGLOG(5, "Waiting for buffer...");
@@ -1678,7 +1734,7 @@ static int ZSTDMT_tryGetInputRange(ZSTDMT_CCtx* mtctx)
typedef struct {
size_t toLoad; /* The number of bytes to load from the input. */
int flush; /* Boolean declaring if we must flush because we found a synchronization point. */
} syncPoint_t;
} SyncPoint;
/**
* Searches through the input for a synchronization point. If one is found, we
@@ -1686,14 +1742,14 @@ typedef struct {
* Otherwise, we will load as many bytes as possible and instruct the caller
* to continue as normal.
*/
static syncPoint_t
static SyncPoint
findSynchronizationPoint(ZSTDMT_CCtx const* mtctx, ZSTD_inBuffer const input)
{
BYTE const* const istart = (BYTE const*)input.src + input.pos;
U64 const primePower = mtctx->rsync.primePower;
U64 const hitMask = mtctx->rsync.hitMask;
syncPoint_t syncPoint;
SyncPoint syncPoint;
U64 hash;
BYTE const* prev;
size_t pos;
@@ -1825,7 +1881,7 @@ size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
DEBUGLOG(5, "ZSTDMT_tryGetInputRange completed successfully : mtctx->inBuff.buffer.start = %p", mtctx->inBuff.buffer.start);
}
if (mtctx->inBuff.buffer.start != NULL) {
syncPoint_t const syncPoint = findSynchronizationPoint(mtctx, *input);
SyncPoint const syncPoint = findSynchronizationPoint(mtctx, *input);
if (syncPoint.flush && endOp == ZSTD_e_continue) {
endOp = ZSTD_e_flush;
}
Utilities/cmzstd/lib/compress/zstdmt_compress.h
+4 -15
View File
@@ -11,10 +11,10 @@
#ifndef ZSTDMT_COMPRESS_H
#define ZSTDMT_COMPRESS_H
#if defined (__cplusplus)
extern "C" {
#endif
/* === Dependencies === */
#include "../common/zstd_deps.h" /* size_t */
#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_parameters */
#include "../zstd.h" /* ZSTD_inBuffer, ZSTD_outBuffer, ZSTDLIB_API */
/* Note : This is an internal API.
* These APIs used to be exposed with ZSTDLIB_API,
@@ -25,12 +25,6 @@
* otherwise ZSTDMT_createCCtx*() will fail.
*/
/* === Dependencies === */
#include "../common/zstd_deps.h" /* size_t */
#define ZSTD_STATIC_LINKING_ONLY /* ZSTD_parameters */
#include "../zstd.h" /* ZSTD_inBuffer, ZSTD_outBuffer, ZSTDLIB_API */
/* === Constants === */
#ifndef ZSTDMT_NBWORKERS_MAX /* a different value can be selected at compile time */
# define ZSTDMT_NBWORKERS_MAX ((sizeof(void*)==4) /*32-bit*/ ? 64 : 256)
@@ -105,9 +99,4 @@ void ZSTDMT_updateCParams_whileCompressing(ZSTDMT_CCtx* mtctx, const ZSTD_CCtx_p
*/
ZSTD_frameProgression ZSTDMT_getFrameProgression(ZSTDMT_CCtx* mtctx);
#if defined (__cplusplus)
}
#endif
#endif /* ZSTDMT_COMPRESS_H */
Utilities/cmzstd/lib/decompress/huf_decompress.c
+188 -126
View File
@@ -34,6 +34,12 @@
* Macros
****************************************************************/
#ifdef HUF_DISABLE_FAST_DECODE
# define HUF_ENABLE_FAST_DECODE 0
#else
# define HUF_ENABLE_FAST_DECODE 1
#endif
/* These two optional macros force the use one way or another of the two
* Huffman decompression implementations. You can't force in both directions
* at the same time.
@@ -158,17 +164,18 @@ static size_t HUF_initFastDStream(BYTE const* ip) {
* op [in/out] - The output pointers, must be updated to reflect what is written.
* bits [in/out] - The bitstream containers, must be updated to reflect the current state.
* dt [in] - The decoding table.
* ilimit [in] - The input limit, stop when any input pointer is below ilimit.
* ilowest [in] - The beginning of the valid range of the input. Decoders may read
* down to this pointer. It may be below iend[0].
* oend [in] - The end of the output stream. op[3] must not cross oend.
* iend [in] - The end of each input stream. ip[i] may cross iend[i],
* as long as it is above ilimit, but that indicates corruption.
* as long as it is above ilowest, but that indicates corruption.
*/
typedef struct {
BYTE const* ip[4];
BYTE* op[4];
U64 bits[4];
void const* dt;
BYTE const* ilimit;
BYTE const* ilowest;
BYTE* oend;
BYTE const* iend[4];
} HUF_DecompressFastArgs;
@@ -186,9 +193,9 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
void const* dt = DTable + 1;
U32 const dtLog = HUF_getDTableDesc(DTable).tableLog;
const BYTE* const ilimit = (const BYTE*)src + 6 + 8;
const BYTE* const istart = (const BYTE*)src;
BYTE* const oend = (BYTE*)dst + dstSize;
BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
/* The fast decoding loop assumes 64-bit little-endian.
* This condition is false on x32.
@@ -196,6 +203,11 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
if (!MEM_isLittleEndian() || MEM_32bits())
return 0;
/* Avoid nullptr addition */
if (dstSize == 0)
return 0;
assert(dst != NULL);
/* strict minimum : jump table + 1 byte per stream */
if (srcSize < 10)
return ERROR(corruption_detected);
@@ -209,7 +221,6 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
/* Read the jump table. */
{
const BYTE* const istart = (const BYTE*)src;
size_t const length1 = MEM_readLE16(istart);
size_t const length2 = MEM_readLE16(istart+2);
size_t const length3 = MEM_readLE16(istart+4);
@@ -221,10 +232,8 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
/* HUF_initFastDStream() requires this, and this small of an input
* won't benefit from the ASM loop anyways.
* length1 must be >= 16 so that ip[0] >= ilimit before the loop
* starts.
*/
if (length1 < 16 || length2 < 8 || length3 < 8 || length4 < 8)
if (length1 < 8 || length2 < 8 || length3 < 8 || length4 < 8)
return 0;
if (length4 > srcSize) return ERROR(corruption_detected); /* overflow */
}
@@ -256,11 +265,12 @@ static size_t HUF_DecompressFastArgs_init(HUF_DecompressFastArgs* args, void* ds
args->bits[2] = HUF_initFastDStream(args->ip[2]);
args->bits[3] = HUF_initFastDStream(args->ip[3]);
/* If ip[] >= ilimit, it is guaranteed to be safe to
* reload bits[]. It may be beyond its section, but is
* guaranteed to be valid (>= istart).
*/
args->ilimit = ilimit;
/* The decoders must be sure to never read beyond ilowest.
* This is lower than iend[0], but allowing decoders to read
* down to ilowest can allow an extra iteration or two in the
* fast loop.
*/
args->ilowest = istart;
args->oend = oend;
args->dt = dt;
@@ -285,13 +295,31 @@ static size_t HUF_initRemainingDStream(BIT_DStream_t* bit, HUF_DecompressFastArg
assert(sizeof(size_t) == 8);
bit->bitContainer = MEM_readLEST(args->ip[stream]);
bit->bitsConsumed = ZSTD_countTrailingZeros64(args->bits[stream]);
bit->start = (const char*)args->iend[0];
bit->start = (const char*)args->ilowest;
bit->limitPtr = bit->start + sizeof(size_t);
bit->ptr = (const char*)args->ip[stream];
return 0;
}
/* Calls X(N) for each stream 0, 1, 2, 3. */
#define HUF_4X_FOR_EACH_STREAM(X) \
do { \
X(0); \
X(1); \
X(2); \
X(3); \
} while (0)
/* Calls X(N, var) for each stream 0, 1, 2, 3. */
#define HUF_4X_FOR_EACH_STREAM_WITH_VAR(X, var) \
do { \
X(0, (var)); \
X(1, (var)); \
X(2, (var)); \
X(3, (var)); \
} while (0)
#ifndef HUF_FORCE_DECOMPRESS_X2
@@ -500,15 +528,19 @@ HUF_decodeSymbolX1(BIT_DStream_t* Dstream, const HUF_DEltX1* dt, const U32 dtLog
}
#define HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr) \
*ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog)
do { *ptr++ = HUF_decodeSymbolX1(DStreamPtr, dt, dtLog); } while (0)
#define HUF_DECODE_SYMBOLX1_1(ptr, DStreamPtr) \
if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
#define HUF_DECODE_SYMBOLX1_1(ptr, DStreamPtr) \
do { \
if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr); \
} while (0)
#define HUF_DECODE_SYMBOLX1_2(ptr, DStreamPtr) \
if (MEM_64bits()) \
HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr)
#define HUF_DECODE_SYMBOLX1_2(ptr, DStreamPtr) \
do { \
if (MEM_64bits()) \
HUF_DECODE_SYMBOLX1_0(ptr, DStreamPtr); \
} while (0)
HINT_INLINE size_t
HUF_decodeStreamX1(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX1* const dt, const U32 dtLog)
@@ -546,7 +578,7 @@ HUF_decompress1X1_usingDTable_internal_body(
const HUF_DTable* DTable)
{
BYTE* op = (BYTE*)dst;
BYTE* const oend = op + dstSize;
BYTE* const oend = ZSTD_maybeNullPtrAdd(op, dstSize);
const void* dtPtr = DTable + 1;
const HUF_DEltX1* const dt = (const HUF_DEltX1*)dtPtr;
BIT_DStream_t bitD;
@@ -574,6 +606,7 @@ HUF_decompress4X1_usingDTable_internal_body(
{
/* Check */
if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
if (dstSize < 6) return ERROR(corruption_detected); /* stream 4-split doesn't work */
{ const BYTE* const istart = (const BYTE*) cSrc;
BYTE* const ostart = (BYTE*) dst;
@@ -609,7 +642,7 @@ HUF_decompress4X1_usingDTable_internal_body(
if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
if (opStart4 > oend) return ERROR(corruption_detected); /* overflow */
if (dstSize < 6) return ERROR(corruption_detected); /* stream 4-split doesn't work */
assert(dstSize >= 6); /* validated above */
CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
@@ -692,7 +725,7 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
BYTE* op[4];
U16 const* const dtable = (U16 const*)args->dt;
BYTE* const oend = args->oend;
BYTE const* const ilimit = args->ilimit;
BYTE const* const ilowest = args->ilowest;
/* Copy the arguments to local variables */
ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
@@ -705,13 +738,12 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
for (;;) {
BYTE* olimit;
int stream;
int symbol;
/* Assert loop preconditions */
#ifndef NDEBUG
for (stream = 0; stream < 4; ++stream) {
assert(op[stream] <= (stream == 3 ? oend : op[stream + 1]));
assert(ip[stream] >= ilimit);
assert(ip[stream] >= ilowest);
}
#endif
/* Compute olimit */
@@ -721,7 +753,7 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
/* Each iteration consumes up to 11 bits * 5 = 55 bits < 7 bytes
* per stream.
*/
size_t const iiters = (size_t)(ip[0] - ilimit) / 7;
size_t const iiters = (size_t)(ip[0] - ilowest) / 7;
/* We can safely run iters iterations before running bounds checks */
size_t const iters = MIN(oiters, iiters);
size_t const symbols = iters * 5;
@@ -732,8 +764,8 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
*/
olimit = op[3] + symbols;
/* Exit fast decoding loop once we get close to the end. */
if (op[3] + 20 > olimit)
/* Exit fast decoding loop once we reach the end. */
if (op[3] == olimit)
break;
/* Exit the decoding loop if any input pointer has crossed the
@@ -752,27 +784,42 @@ void HUF_decompress4X1_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
}
#endif
#define HUF_4X1_DECODE_SYMBOL(_stream, _symbol) \
do { \
int const index = (int)(bits[(_stream)] >> 53); \
int const entry = (int)dtable[index]; \
bits[(_stream)] <<= (entry & 0x3F); \
op[(_stream)][(_symbol)] = (BYTE)((entry >> 8) & 0xFF); \
} while (0)
#define HUF_4X1_RELOAD_STREAM(_stream) \
do { \
int const ctz = ZSTD_countTrailingZeros64(bits[(_stream)]); \
int const nbBits = ctz & 7; \
int const nbBytes = ctz >> 3; \
op[(_stream)] += 5; \
ip[(_stream)] -= nbBytes; \
bits[(_stream)] = MEM_read64(ip[(_stream)]) | 1; \
bits[(_stream)] <<= nbBits; \
} while (0)
/* Manually unroll the loop because compilers don't consistently
* unroll the inner loops, which destroys performance.
*/
do {
/* Decode 5 symbols in each of the 4 streams */
for (symbol = 0; symbol < 5; ++symbol) {
for (stream = 0; stream < 4; ++stream) {
int const index = (int)(bits[stream] >> 53);
int const entry = (int)dtable[index];
bits[stream] <<= (entry & 63);
op[stream][symbol] = (BYTE)((entry >> 8) & 0xFF);
}
}
/* Reload the bitstreams */
for (stream = 0; stream < 4; ++stream) {
int const ctz = ZSTD_countTrailingZeros64(bits[stream]);
int const nbBits = ctz & 7;
int const nbBytes = ctz >> 3;
op[stream] += 5;
ip[stream] -= nbBytes;
bits[stream] = MEM_read64(ip[stream]) | 1;
bits[stream] <<= nbBits;
}
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 0);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 1);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 2);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 3);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X1_DECODE_SYMBOL, 4);
/* Reload each of the 4 the bitstreams */
HUF_4X_FOR_EACH_STREAM(HUF_4X1_RELOAD_STREAM);
} while (op[3] < olimit);
#undef HUF_4X1_DECODE_SYMBOL
#undef HUF_4X1_RELOAD_STREAM
}
_out:
@@ -797,8 +844,8 @@ HUF_decompress4X1_usingDTable_internal_fast(
HUF_DecompressFastLoopFn loopFn)
{
void const* dt = DTable + 1;
const BYTE* const iend = (const BYTE*)cSrc + 6;
BYTE* const oend = (BYTE*)dst + dstSize;
BYTE const* const ilowest = (BYTE const*)cSrc;
BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
HUF_DecompressFastArgs args;
{ size_t const ret = HUF_DecompressFastArgs_init(&args, dst, dstSize, cSrc, cSrcSize, DTable);
FORWARD_IF_ERROR(ret, "Failed to init fast loop args");
@@ -806,18 +853,22 @@ HUF_decompress4X1_usingDTable_internal_fast(
return 0;
}
assert(args.ip[0] >= args.ilimit);
assert(args.ip[0] >= args.ilowest);
loopFn(&args);
/* Our loop guarantees that ip[] >= ilimit and that we haven't
/* Our loop guarantees that ip[] >= ilowest and that we haven't
* overwritten any op[].
*/
assert(args.ip[0] >= iend);
assert(args.ip[1] >= iend);
assert(args.ip[2] >= iend);
assert(args.ip[3] >= iend);
assert(args.ip[0] >= ilowest);
assert(args.ip[0] >= ilowest);
assert(args.ip[1] >= ilowest);
assert(args.ip[2] >= ilowest);
assert(args.ip[3] >= ilowest);
assert(args.op[3] <= oend);
(void)iend;
assert(ilowest == args.ilowest);
assert(ilowest + 6 == args.iend[0]);
(void)ilowest;
/* finish bit streams one by one. */
{ size_t const segmentSize = (dstSize+3) / 4;
@@ -868,7 +919,7 @@ static size_t HUF_decompress4X1_usingDTable_internal(void* dst, size_t dstSize,
}
#endif
if (!(flags & HUF_flags_disableFast)) {
if (HUF_ENABLE_FAST_DECODE && !(flags & HUF_flags_disableFast)) {
size_t const ret = HUF_decompress4X1_usingDTable_internal_fast(dst, dstSize, cSrc, cSrcSize, DTable, loopFn);
if (ret != 0)
return ret;
@@ -1239,15 +1290,19 @@ HUF_decodeLastSymbolX2(void* op, BIT_DStream_t* DStream, const HUF_DEltX2* dt, c
}
#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
do { ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); } while (0)
#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
do { \
if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); \
} while (0)
#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
if (MEM_64bits()) \
ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog)
#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
do { \
if (MEM_64bits()) \
ptr += HUF_decodeSymbolX2(ptr, DStreamPtr, dt, dtLog); \
} while (0)
HINT_INLINE size_t
HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd,
@@ -1307,7 +1362,7 @@ HUF_decompress1X2_usingDTable_internal_body(
/* decode */
{ BYTE* const ostart = (BYTE*) dst;
BYTE* const oend = ostart + dstSize;
BYTE* const oend = ZSTD_maybeNullPtrAdd(ostart, dstSize);
const void* const dtPtr = DTable+1; /* force compiler to not use strict-aliasing */
const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
DTableDesc const dtd = HUF_getDTableDesc(DTable);
@@ -1332,6 +1387,7 @@ HUF_decompress4X2_usingDTable_internal_body(
const HUF_DTable* DTable)
{
if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
if (dstSize < 6) return ERROR(corruption_detected); /* stream 4-split doesn't work */
{ const BYTE* const istart = (const BYTE*) cSrc;
BYTE* const ostart = (BYTE*) dst;
@@ -1367,7 +1423,7 @@ HUF_decompress4X2_usingDTable_internal_body(
if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
if (opStart4 > oend) return ERROR(corruption_detected); /* overflow */
if (dstSize < 6) return ERROR(corruption_detected); /* stream 4-split doesn't work */
assert(dstSize >= 6 /* validated above */);
CHECK_F( BIT_initDStream(&bitD1, istart1, length1) );
CHECK_F( BIT_initDStream(&bitD2, istart2, length2) );
CHECK_F( BIT_initDStream(&bitD3, istart3, length3) );
@@ -1472,7 +1528,7 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
BYTE* op[4];
BYTE* oend[4];
HUF_DEltX2 const* const dtable = (HUF_DEltX2 const*)args->dt;
BYTE const* const ilimit = args->ilimit;
BYTE const* const ilowest = args->ilowest;
/* Copy the arguments to local registers. */
ZSTD_memcpy(&bits, &args->bits, sizeof(bits));
@@ -1490,13 +1546,12 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
for (;;) {
BYTE* olimit;
int stream;
int symbol;
/* Assert loop preconditions */
#ifndef NDEBUG
for (stream = 0; stream < 4; ++stream) {
assert(op[stream] <= oend[stream]);
assert(ip[stream] >= ilimit);
assert(ip[stream] >= ilowest);
}
#endif
/* Compute olimit */
@@ -1509,7 +1564,7 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
* We also know that each input pointer is >= ip[0]. So we can run
* iters loops before running out of input.
*/
size_t iters = (size_t)(ip[0] - ilimit) / 7;
size_t iters = (size_t)(ip[0] - ilowest) / 7;
/* Each iteration can produce up to 10 bytes of output per stream.
* Each output stream my advance at different rates. So take the
* minimum number of safe iterations among all the output streams.
@@ -1527,8 +1582,8 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
*/
olimit = op[3] + (iters * 5);
/* Exit the fast decoding loop if we are too close to the end. */
if (op[3] + 10 > olimit)
/* Exit the fast decoding loop once we reach the end. */
if (op[3] == olimit)
break;
/* Exit the decoding loop if any input pointer has crossed the
@@ -1547,54 +1602,58 @@ void HUF_decompress4X2_usingDTable_internal_fast_c_loop(HUF_DecompressFastArgs*
}
#endif
#define HUF_4X2_DECODE_SYMBOL(_stream, _decode3) \
do { \
if ((_decode3) || (_stream) != 3) { \
int const index = (int)(bits[(_stream)] >> 53); \
HUF_DEltX2 const entry = dtable[index]; \
MEM_write16(op[(_stream)], entry.sequence); \
bits[(_stream)] <<= (entry.nbBits) & 0x3F; \
op[(_stream)] += (entry.length); \
} \
} while (0)
#define HUF_4X2_RELOAD_STREAM(_stream) \
do { \
HUF_4X2_DECODE_SYMBOL(3, 1); \
{ \
int const ctz = ZSTD_countTrailingZeros64(bits[(_stream)]); \
int const nbBits = ctz & 7; \
int const nbBytes = ctz >> 3; \
ip[(_stream)] -= nbBytes; \
bits[(_stream)] = MEM_read64(ip[(_stream)]) | 1; \
bits[(_stream)] <<= nbBits; \
} \
} while (0)
/* Manually unroll the loop because compilers don't consistently
* unroll the inner loops, which destroys performance.
*/
do {
/* Do 5 table lookups for each of the first 3 streams */
for (symbol = 0; symbol < 5; ++symbol) {
for (stream = 0; stream < 3; ++stream) {
int const index = (int)(bits[stream] >> 53);
HUF_DEltX2 const entry = dtable[index];
MEM_write16(op[stream], entry.sequence);
bits[stream] <<= (entry.nbBits);
op[stream] += (entry.length);
}
}
/* Do 1 table lookup from the final stream */
{
int const index = (int)(bits[3] >> 53);
HUF_DEltX2 const entry = dtable[index];
MEM_write16(op[3], entry.sequence);
bits[3] <<= (entry.nbBits);
op[3] += (entry.length);
}
/* Do 4 table lookups from the final stream & reload bitstreams */
for (stream = 0; stream < 4; ++stream) {
/* Do a table lookup from the final stream.
* This is interleaved with the reloading to reduce register
* pressure. This shouldn't be necessary, but compilers can
* struggle with codegen with high register pressure.
*/
{
int const index = (int)(bits[3] >> 53);
HUF_DEltX2 const entry = dtable[index];
MEM_write16(op[3], entry.sequence);
bits[3] <<= (entry.nbBits);
op[3] += (entry.length);
}
/* Reload the bistreams. The final bitstream must be reloaded
* after the 5th symbol was decoded.
*/
{
int const ctz = ZSTD_countTrailingZeros64(bits[stream]);
int const nbBits = ctz & 7;
int const nbBytes = ctz >> 3;
ip[stream] -= nbBytes;
bits[stream] = MEM_read64(ip[stream]) | 1;
bits[stream] <<= nbBits;
}
}
/* Decode 5 symbols from each of the first 3 streams.
* The final stream will be decoded during the reload phase
* to reduce register pressure.
*/
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
HUF_4X_FOR_EACH_STREAM_WITH_VAR(HUF_4X2_DECODE_SYMBOL, 0);
/* Decode one symbol from the final stream */
HUF_4X2_DECODE_SYMBOL(3, 1);
/* Decode 4 symbols from the final stream & reload bitstreams.
* The final stream is reloaded last, meaning that all 5 symbols
* are decoded from the final stream before it is reloaded.
*/
HUF_4X_FOR_EACH_STREAM(HUF_4X2_RELOAD_STREAM);
} while (op[3] < olimit);
}
#undef HUF_4X2_DECODE_SYMBOL
#undef HUF_4X2_RELOAD_STREAM
_out:
/* Save the final values of each of the state variables back to args. */
@@ -1611,8 +1670,8 @@ HUF_decompress4X2_usingDTable_internal_fast(
const HUF_DTable* DTable,
HUF_DecompressFastLoopFn loopFn) {
void const* dt = DTable + 1;
const BYTE* const iend = (const BYTE*)cSrc + 6;
BYTE* const oend = (BYTE*)dst + dstSize;
const BYTE* const ilowest = (const BYTE*)cSrc;
BYTE* const oend = ZSTD_maybeNullPtrAdd((BYTE*)dst, dstSize);
HUF_DecompressFastArgs args;
{
size_t const ret = HUF_DecompressFastArgs_init(&args, dst, dstSize, cSrc, cSrcSize, DTable);
@@ -1621,16 +1680,19 @@ HUF_decompress4X2_usingDTable_internal_fast(
return 0;
}
assert(args.ip[0] >= args.ilimit);
assert(args.ip[0] >= args.ilowest);
loopFn(&args);
/* note : op4 already verified within main loop */
assert(args.ip[0] >= iend);
assert(args.ip[1] >= iend);
assert(args.ip[2] >= iend);
assert(args.ip[3] >= iend);
assert(args.ip[0] >= ilowest);
assert(args.ip[1] >= ilowest);
assert(args.ip[2] >= ilowest);
assert(args.ip[3] >= ilowest);
assert(args.op[3] <= oend);
(void)iend;
assert(ilowest == args.ilowest);
assert(ilowest + 6 == args.iend[0]);
(void)ilowest;
/* finish bitStreams one by one */
{
@@ -1679,7 +1741,7 @@ static size_t HUF_decompress4X2_usingDTable_internal(void* dst, size_t dstSize,
}
#endif
if (!(flags & HUF_flags_disableFast)) {
if (HUF_ENABLE_FAST_DECODE && !(flags & HUF_flags_disableFast)) {
size_t const ret = HUF_decompress4X2_usingDTable_internal_fast(dst, dstSize, cSrc, cSrcSize, DTable, loopFn);
if (ret != 0)
return ret;
Utilities/cmzstd/lib/decompress/zstd_decompress.c
+98 -43
View File
@@ -55,18 +55,19 @@
/*-*******************************************************
* Dependencies
*********************************************************/
#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
#include "../common/zstd_deps.h" /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
#include "../common/allocations.h" /* ZSTD_customMalloc, ZSTD_customCalloc, ZSTD_customFree */
#include "../common/error_private.h"
#include "../common/zstd_internal.h" /* blockProperties_t */
#include "../common/mem.h" /* low level memory routines */
#include "../common/bits.h" /* ZSTD_highbit32 */
#define FSE_STATIC_LINKING_ONLY
#include "../common/fse.h"
#include "../common/huf.h"
#include "../common/xxhash.h" /* XXH64_reset, XXH64_update, XXH64_digest, XXH64 */
#include "../common/zstd_internal.h" /* blockProperties_t */
#include "zstd_decompress_internal.h" /* ZSTD_DCtx */
#include "zstd_ddict.h" /* ZSTD_DDictDictContent */
#include "zstd_decompress_block.h" /* ZSTD_decompressBlock_internal */
#include "../common/bits.h" /* ZSTD_highbit32 */
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
# include "../legacy/zstd_legacy.h"
@@ -245,6 +246,7 @@ static void ZSTD_DCtx_resetParameters(ZSTD_DCtx* dctx)
dctx->forceIgnoreChecksum = ZSTD_d_validateChecksum;
dctx->refMultipleDDicts = ZSTD_rmd_refSingleDDict;
dctx->disableHufAsm = 0;
dctx->maxBlockSizeParam = 0;
}
static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
@@ -265,6 +267,7 @@ static void ZSTD_initDCtx_internal(ZSTD_DCtx* dctx)
#endif
dctx->noForwardProgress = 0;
dctx->oversizedDuration = 0;
dctx->isFrameDecompression = 1;
#if DYNAMIC_BMI2
dctx->bmi2 = ZSTD_cpuSupportsBmi2();
#endif
@@ -441,7 +444,7 @@ size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize)
* @return : 0, `zfhPtr` is correctly filled,
* >0, `srcSize` is too small, value is wanted `srcSize` amount,
** or an error code, which can be tested using ZSTD_isError() */
size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)
size_t ZSTD_getFrameHeader_advanced(ZSTD_FrameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format)
{
const BYTE* ip = (const BYTE*)src;
size_t const minInputSize = ZSTD_startingInputLength(format);
@@ -481,8 +484,10 @@ size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, s
if (srcSize < ZSTD_SKIPPABLEHEADERSIZE)
return ZSTD_SKIPPABLEHEADERSIZE; /* magic number + frame length */
ZSTD_memset(zfhPtr, 0, sizeof(*zfhPtr));
zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE);
zfhPtr->frameType = ZSTD_skippableFrame;
zfhPtr->dictID = MEM_readLE32(src) - ZSTD_MAGIC_SKIPPABLE_START;
zfhPtr->headerSize = ZSTD_SKIPPABLEHEADERSIZE;
zfhPtr->frameContentSize = MEM_readLE32((const char *)src + ZSTD_FRAMEIDSIZE);
return 0;
}
RETURN_ERROR(prefix_unknown, "");
@@ -551,7 +556,7 @@ size_t ZSTD_getFrameHeader_advanced(ZSTD_frameHeader* zfhPtr, const void* src, s
* @return : 0, `zfhPtr` is correctly filled,
* >0, `srcSize` is too small, value is wanted `srcSize` amount,
* or an error code, which can be tested using ZSTD_isError() */
size_t ZSTD_getFrameHeader(ZSTD_frameHeader* zfhPtr, const void* src, size_t srcSize)
size_t ZSTD_getFrameHeader(ZSTD_FrameHeader* zfhPtr, const void* src, size_t srcSize)
{
return ZSTD_getFrameHeader_advanced(zfhPtr, src, srcSize, ZSTD_f_zstd1);
}
@@ -569,7 +574,7 @@ unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize)
return ret == 0 ? ZSTD_CONTENTSIZE_UNKNOWN : ret;
}
#endif
{ ZSTD_frameHeader zfh;
{ ZSTD_FrameHeader zfh;
if (ZSTD_getFrameHeader(&zfh, src, srcSize) != 0)
return ZSTD_CONTENTSIZE_ERROR;
if (zfh.frameType == ZSTD_skippableFrame) {
@@ -726,17 +731,17 @@ static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)
return frameSizeInfo;
}
static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)
static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize, ZSTD_format_e format)
{
ZSTD_frameSizeInfo frameSizeInfo;
ZSTD_memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
if (ZSTD_isLegacy(src, srcSize))
if (format == ZSTD_f_zstd1 && ZSTD_isLegacy(src, srcSize))
return ZSTD_findFrameSizeInfoLegacy(src, srcSize);
#endif
if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
if (format == ZSTD_f_zstd1 && (srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
&& (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);
assert(ZSTD_isError(frameSizeInfo.compressedSize) ||
@@ -747,10 +752,10 @@ static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize
const BYTE* const ipstart = ip;
size_t remainingSize = srcSize;
size_t nbBlocks = 0;
ZSTD_frameHeader zfh;
ZSTD_FrameHeader zfh;
/* Extract Frame Header */
{ size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
{ size_t const ret = ZSTD_getFrameHeader_advanced(&zfh, src, srcSize, format);
if (ZSTD_isError(ret))
return ZSTD_errorFrameSizeInfo(ret);
if (ret > 0)
@@ -793,20 +798,22 @@ static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize
}
}
static size_t ZSTD_findFrameCompressedSize_advanced(const void *src, size_t srcSize, ZSTD_format_e format) {
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, format);
return frameSizeInfo.compressedSize;
}
/** ZSTD_findFrameCompressedSize() :
* compatible with legacy mode
* `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame
* `srcSize` must be at least as large as the frame contained
* @return : the compressed size of the frame starting at `src` */
* See docs in zstd.h
* Note: compatible with legacy mode */
size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
{
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
return frameSizeInfo.compressedSize;
return ZSTD_findFrameCompressedSize_advanced(src, srcSize, ZSTD_f_zstd1);
}
/** ZSTD_decompressBound() :
* compatible with legacy mode
* `src` must point to the start of a ZSTD frame or a skippeable frame
* `src` must point to the start of a ZSTD frame or a skippable frame
* `srcSize` must be at least as large as the frame contained
* @return : the maximum decompressed size of the compressed source
*/
@@ -815,7 +822,7 @@ unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize)
unsigned long long bound = 0;
/* Iterate over each frame */
while (srcSize > 0) {
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, ZSTD_f_zstd1);
size_t const compressedSize = frameSizeInfo.compressedSize;
unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)
@@ -835,10 +842,10 @@ size_t ZSTD_decompressionMargin(void const* src, size_t srcSize)
/* Iterate over each frame */
while (srcSize > 0) {
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize, ZSTD_f_zstd1);
size_t const compressedSize = frameSizeInfo.compressedSize;
unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
ZSTD_frameHeader zfh;
ZSTD_FrameHeader zfh;
FORWARD_IF_ERROR(ZSTD_getFrameHeader(&zfh, src, srcSize), "");
if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)
@@ -912,7 +919,7 @@ static size_t ZSTD_setRleBlock(void* dst, size_t dstCapacity,
return regenSize;
}
static void ZSTD_DCtx_trace_end(ZSTD_DCtx const* dctx, U64 uncompressedSize, U64 compressedSize, unsigned streaming)
static void ZSTD_DCtx_trace_end(ZSTD_DCtx const* dctx, U64 uncompressedSize, U64 compressedSize, int streaming)
{
#if ZSTD_TRACE
if (dctx->traceCtx && ZSTD_trace_decompress_end != NULL) {
@@ -971,6 +978,10 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;
}
/* Shrink the blockSizeMax if enabled */
if (dctx->maxBlockSizeParam != 0)
dctx->fParams.blockSizeMax = MIN(dctx->fParams.blockSizeMax, (unsigned)dctx->maxBlockSizeParam);
/* Loop on each block */
while (1) {
BYTE* oBlockEnd = oend;
@@ -1003,7 +1014,8 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
switch(blockProperties.blockType)
{
case bt_compressed:
decodedSize = ZSTD_decompressBlock_internal(dctx, op, (size_t)(oBlockEnd-op), ip, cBlockSize, /* frame */ 1, not_streaming);
assert(dctx->isFrameDecompression == 1);
decodedSize = ZSTD_decompressBlock_internal(dctx, op, (size_t)(oBlockEnd-op), ip, cBlockSize, not_streaming);
break;
case bt_raw :
/* Use oend instead of oBlockEnd because this function is safe to overlap. It uses memmove. */
@@ -1016,12 +1028,14 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
default:
RETURN_ERROR(corruption_detected, "invalid block type");
}
if (ZSTD_isError(decodedSize)) return decodedSize;
if (dctx->validateChecksum)
FORWARD_IF_ERROR(decodedSize, "Block decompression failure");
DEBUGLOG(5, "Decompressed block of dSize = %u", (unsigned)decodedSize);
if (dctx->validateChecksum) {
XXH64_update(&dctx->xxhState, op, decodedSize);
if (decodedSize != 0)
}
if (decodedSize) /* support dst = NULL,0 */ {
op += decodedSize;
}
assert(ip != NULL);
ip += cBlockSize;
remainingSrcSize -= cBlockSize;
@@ -1045,13 +1059,15 @@ static size_t ZSTD_decompressFrame(ZSTD_DCtx* dctx,
}
ZSTD_DCtx_trace_end(dctx, (U64)(op-ostart), (U64)(ip-istart), /* streaming */ 0);
/* Allow caller to get size read */
DEBUGLOG(4, "ZSTD_decompressFrame: decompressed frame of size %zi, consuming %zi bytes of input", op-ostart, ip - (const BYTE*)*srcPtr);
DEBUGLOG(4, "ZSTD_decompressFrame: decompressed frame of size %i, consuming %i bytes of input", (int)(op-ostart), (int)(ip - (const BYTE*)*srcPtr));
*srcPtr = ip;
*srcSizePtr = remainingSrcSize;
return (size_t)(op-ostart);
}
static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
static
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
const void* dict, size_t dictSize,
@@ -1071,7 +1087,7 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
while (srcSize >= ZSTD_startingInputLength(dctx->format)) {
#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
if (ZSTD_isLegacy(src, srcSize)) {
if (dctx->format == ZSTD_f_zstd1 && ZSTD_isLegacy(src, srcSize)) {
size_t decodedSize;
size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize);
if (ZSTD_isError(frameSize)) return frameSize;
@@ -1081,6 +1097,15 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize);
if (ZSTD_isError(decodedSize)) return decodedSize;
{
unsigned long long const expectedSize = ZSTD_getFrameContentSize(src, srcSize);
RETURN_ERROR_IF(expectedSize == ZSTD_CONTENTSIZE_ERROR, corruption_detected, "Corrupted frame header!");
if (expectedSize != ZSTD_CONTENTSIZE_UNKNOWN) {
RETURN_ERROR_IF(expectedSize != decodedSize, corruption_detected,
"Frame header size does not match decoded size!");
}
}
assert(decodedSize <= dstCapacity);
dst = (BYTE*)dst + decodedSize;
dstCapacity -= decodedSize;
@@ -1092,7 +1117,7 @@ static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx* dctx,
}
#endif
if (srcSize >= 4) {
if (dctx->format == ZSTD_f_zstd1 && srcSize >= 4) {
U32 const magicNumber = MEM_readLE32(src);
DEBUGLOG(5, "reading magic number %08X", (unsigned)magicNumber);
if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
@@ -1319,7 +1344,8 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
{
case bt_compressed:
DEBUGLOG(5, "ZSTD_decompressContinue: case bt_compressed");
rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 1, is_streaming);
assert(dctx->isFrameDecompression == 1);
rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, is_streaming);
dctx->expected = 0; /* Streaming not supported */
break;
case bt_raw :
@@ -1388,6 +1414,7 @@ size_t ZSTD_decompressContinue(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, c
case ZSTDds_decodeSkippableHeader:
assert(src != NULL);
assert(srcSize <= ZSTD_SKIPPABLEHEADERSIZE);
assert(dctx->format != ZSTD_f_zstd1_magicless);
ZSTD_memcpy(dctx->headerBuffer + (ZSTD_SKIPPABLEHEADERSIZE - srcSize), src, srcSize); /* complete skippable header */
dctx->expected = MEM_readLE32(dctx->headerBuffer + ZSTD_FRAMEIDSIZE); /* note : dctx->expected can grow seriously large, beyond local buffer size */
dctx->stage = ZSTDds_skipFrame;
@@ -1548,6 +1575,7 @@ size_t ZSTD_decompressBegin(ZSTD_DCtx* dctx)
dctx->litEntropy = dctx->fseEntropy = 0;
dctx->dictID = 0;
dctx->bType = bt_reserved;
dctx->isFrameDecompression = 1;
ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
ZSTD_memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */
dctx->LLTptr = dctx->entropy.LLTable;
@@ -1615,7 +1643,7 @@ unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize)
* ZSTD_getFrameHeader(), which will provide a more precise error code. */
unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize)
{
ZSTD_frameHeader zfp = { 0, 0, 0, ZSTD_frame, 0, 0, 0, 0, 0 };
ZSTD_FrameHeader zfp = { 0, 0, 0, ZSTD_frame, 0, 0, 0, 0, 0 };
size_t const hError = ZSTD_getFrameHeader(&zfp, src, srcSize);
if (ZSTD_isError(hError)) return 0;
return zfp.dictID;
@@ -1819,6 +1847,10 @@ ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam)
bounds.lowerBound = 0;
bounds.upperBound = 1;
return bounds;
case ZSTD_d_maxBlockSize:
bounds.lowerBound = ZSTD_BLOCKSIZE_MAX_MIN;
bounds.upperBound = ZSTD_BLOCKSIZE_MAX;
return bounds;
default:;
}
@@ -1863,6 +1895,9 @@ size_t ZSTD_DCtx_getParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int* value
case ZSTD_d_disableHuffmanAssembly:
*value = (int)dctx->disableHufAsm;
return 0;
case ZSTD_d_maxBlockSize:
*value = dctx->maxBlockSizeParam;
return 0;
default:;
}
RETURN_ERROR(parameter_unsupported, "");
@@ -1900,6 +1935,10 @@ size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value
CHECK_DBOUNDS(ZSTD_d_disableHuffmanAssembly, value);
dctx->disableHufAsm = value != 0;
return 0;
case ZSTD_d_maxBlockSize:
if (value != 0) CHECK_DBOUNDS(ZSTD_d_maxBlockSize, value);
dctx->maxBlockSizeParam = value;
return 0;
default:;
}
RETURN_ERROR(parameter_unsupported, "");
@@ -1911,6 +1950,7 @@ size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)
|| (reset == ZSTD_reset_session_and_parameters) ) {
dctx->streamStage = zdss_init;
dctx->noForwardProgress = 0;
dctx->isFrameDecompression = 1;
}
if ( (reset == ZSTD_reset_parameters)
|| (reset == ZSTD_reset_session_and_parameters) ) {
@@ -1927,11 +1967,17 @@ size_t ZSTD_sizeof_DStream(const ZSTD_DStream* dctx)
return ZSTD_sizeof_DCtx(dctx);
}
size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
static size_t ZSTD_decodingBufferSize_internal(unsigned long long windowSize, unsigned long long frameContentSize, size_t blockSizeMax)
{
size_t const blockSize = (size_t) MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
/* space is needed to store the litbuffer after the output of a given block without stomping the extDict of a previous run, as well as to cover both windows against wildcopy*/
unsigned long long const neededRBSize = windowSize + blockSize + ZSTD_BLOCKSIZE_MAX + (WILDCOPY_OVERLENGTH * 2);
size_t const blockSize = MIN((size_t)MIN(windowSize, ZSTD_BLOCKSIZE_MAX), blockSizeMax);
/* We need blockSize + WILDCOPY_OVERLENGTH worth of buffer so that if a block
* ends at windowSize + WILDCOPY_OVERLENGTH + 1 bytes, we can start writing
* the block at the beginning of the output buffer, and maintain a full window.
*
* We need another blockSize worth of buffer so that we can store split
* literals at the end of the block without overwriting the extDict window.
*/
unsigned long long const neededRBSize = windowSize + (blockSize * 2) + (WILDCOPY_OVERLENGTH * 2);
unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
size_t const minRBSize = (size_t) neededSize;
RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,
@@ -1939,6 +1985,11 @@ size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long
return minRBSize;
}
size_t ZSTD_decodingBufferSize_min(unsigned long long windowSize, unsigned long long frameContentSize)
{
return ZSTD_decodingBufferSize_internal(windowSize, frameContentSize, ZSTD_BLOCKSIZE_MAX);
}
size_t ZSTD_estimateDStreamSize(size_t windowSize)
{
size_t const blockSize = MIN(windowSize, ZSTD_BLOCKSIZE_MAX);
@@ -1950,7 +2001,7 @@ size_t ZSTD_estimateDStreamSize(size_t windowSize)
size_t ZSTD_estimateDStreamSize_fromFrame(const void* src, size_t srcSize)
{
U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX; /* note : should be user-selectable, but requires an additional parameter (or a dctx) */
ZSTD_frameHeader zfh;
ZSTD_FrameHeader zfh;
size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);
if (ZSTD_isError(err)) return err;
RETURN_ERROR_IF(err>0, srcSize_wrong, "");
@@ -2045,6 +2096,7 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
U32 someMoreWork = 1;
DEBUGLOG(5, "ZSTD_decompressStream");
assert(zds != NULL);
RETURN_ERROR_IF(
input->pos > input->size,
srcSize_wrong,
@@ -2134,12 +2186,12 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
if (zds->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
&& zds->fParams.frameType != ZSTD_skippableFrame
&& (U64)(size_t)(oend-op) >= zds->fParams.frameContentSize) {
size_t const cSize = ZSTD_findFrameCompressedSize(istart, (size_t)(iend-istart));
size_t const cSize = ZSTD_findFrameCompressedSize_advanced(istart, (size_t)(iend-istart), zds->format);
if (cSize <= (size_t)(iend-istart)) {
/* shortcut : using single-pass mode */
size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, (size_t)(oend-op), istart, cSize, ZSTD_getDDict(zds));
if (ZSTD_isError(decompressedSize)) return decompressedSize;
DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")
DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()");
assert(istart != NULL);
ip = istart + cSize;
op = op ? op + decompressedSize : op; /* can occur if frameContentSize = 0 (empty frame) */
@@ -2161,7 +2213,8 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
DEBUGLOG(4, "Consume header");
FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)), "");
if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
if (zds->format == ZSTD_f_zstd1
&& (MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);
zds->stage = ZSTDds_skipFrame;
} else {
@@ -2177,11 +2230,13 @@ size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inB
zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,
frameParameter_windowTooLarge, "");
if (zds->maxBlockSizeParam != 0)
zds->fParams.blockSizeMax = MIN(zds->fParams.blockSizeMax, (unsigned)zds->maxBlockSizeParam);
/* Adapt buffer sizes to frame header instructions */
{ size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
size_t const neededOutBuffSize = zds->outBufferMode == ZSTD_bm_buffered
? ZSTD_decodingBufferSize_min(zds->fParams.windowSize, zds->fParams.frameContentSize)
? ZSTD_decodingBufferSize_internal(zds->fParams.windowSize, zds->fParams.frameContentSize, zds->fParams.blockSizeMax)
: 0;
ZSTD_DCtx_updateOversizedDuration(zds, neededInBuffSize, neededOutBuffSize);
Utilities/cmzstd/lib/decompress/zstd_decompress_block.c
+236 -219
View File
@@ -51,6 +51,13 @@ static void ZSTD_copy4(void* dst, const void* src) { ZSTD_memcpy(dst, src, 4); }
* Block decoding
***************************************************************/
static size_t ZSTD_blockSizeMax(ZSTD_DCtx const* dctx)
{
size_t const blockSizeMax = dctx->isFrameDecompression ? dctx->fParams.blockSizeMax : ZSTD_BLOCKSIZE_MAX;
assert(blockSizeMax <= ZSTD_BLOCKSIZE_MAX);
return blockSizeMax;
}
/*! ZSTD_getcBlockSize() :
* Provides the size of compressed block from block header `src` */
size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
@@ -73,41 +80,49 @@ size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
static void ZSTD_allocateLiteralsBuffer(ZSTD_DCtx* dctx, void* const dst, const size_t dstCapacity, const size_t litSize,
const streaming_operation streaming, const size_t expectedWriteSize, const unsigned splitImmediately)
{
if (streaming == not_streaming && dstCapacity > ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH + litSize + WILDCOPY_OVERLENGTH)
{
/* room for litbuffer to fit without read faulting */
dctx->litBuffer = (BYTE*)dst + ZSTD_BLOCKSIZE_MAX + WILDCOPY_OVERLENGTH;
size_t const blockSizeMax = ZSTD_blockSizeMax(dctx);
assert(litSize <= blockSizeMax);
assert(dctx->isFrameDecompression || streaming == not_streaming);
assert(expectedWriteSize <= blockSizeMax);
if (streaming == not_streaming && dstCapacity > blockSizeMax + WILDCOPY_OVERLENGTH + litSize + WILDCOPY_OVERLENGTH) {
/* If we aren't streaming, we can just put the literals after the output
* of the current block. We don't need to worry about overwriting the
* extDict of our window, because it doesn't exist.
* So if we have space after the end of the block, just put it there.
*/
dctx->litBuffer = (BYTE*)dst + blockSizeMax + WILDCOPY_OVERLENGTH;
dctx->litBufferEnd = dctx->litBuffer + litSize;
dctx->litBufferLocation = ZSTD_in_dst;
}
else if (litSize > ZSTD_LITBUFFEREXTRASIZE)
{
/* won't fit in litExtraBuffer, so it will be split between end of dst and extra buffer */
} else if (litSize <= ZSTD_LITBUFFEREXTRASIZE) {
/* Literals fit entirely within the extra buffer, put them there to avoid
* having to split the literals.
*/
dctx->litBuffer = dctx->litExtraBuffer;
dctx->litBufferEnd = dctx->litBuffer + litSize;
dctx->litBufferLocation = ZSTD_not_in_dst;
} else {
assert(blockSizeMax > ZSTD_LITBUFFEREXTRASIZE);
/* Literals must be split between the output block and the extra lit
* buffer. We fill the extra lit buffer with the tail of the literals,
* and put the rest of the literals at the end of the block, with
* WILDCOPY_OVERLENGTH of buffer room to allow for overreads.
* This MUST not write more than our maxBlockSize beyond dst, because in
* streaming mode, that could overwrite part of our extDict window.
*/
if (splitImmediately) {
/* won't fit in litExtraBuffer, so it will be split between end of dst and extra buffer */
dctx->litBuffer = (BYTE*)dst + expectedWriteSize - litSize + ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH;
dctx->litBufferEnd = dctx->litBuffer + litSize - ZSTD_LITBUFFEREXTRASIZE;
}
else {
} else {
/* initially this will be stored entirely in dst during huffman decoding, it will partially be shifted to litExtraBuffer after */
dctx->litBuffer = (BYTE*)dst + expectedWriteSize - litSize;
dctx->litBufferEnd = (BYTE*)dst + expectedWriteSize;
}
dctx->litBufferLocation = ZSTD_split;
}
else
{
/* fits entirely within litExtraBuffer, so no split is necessary */
dctx->litBuffer = dctx->litExtraBuffer;
dctx->litBufferEnd = dctx->litBuffer + litSize;
dctx->litBufferLocation = ZSTD_not_in_dst;
assert(dctx->litBufferEnd <= (BYTE*)dst + expectedWriteSize);
}
}
/* Hidden declaration for fullbench */
size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
const void* src, size_t srcSize,
void* dst, size_t dstCapacity, const streaming_operation streaming);
/*! ZSTD_decodeLiteralsBlock() :
* Where it is possible to do so without being stomped by the output during decompression, the literals block will be stored
* in the dstBuffer. If there is room to do so, it will be stored in full in the excess dst space after where the current
@@ -116,7 +131,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
*
* @return : nb of bytes read from src (< srcSize )
* note : symbol not declared but exposed for fullbench */
size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
static size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
const void* src, size_t srcSize, /* note : srcSize < BLOCKSIZE */
void* dst, size_t dstCapacity, const streaming_operation streaming)
{
@@ -124,7 +139,8 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
RETURN_ERROR_IF(srcSize < MIN_CBLOCK_SIZE, corruption_detected, "");
{ const BYTE* const istart = (const BYTE*) src;
symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
SymbolEncodingType_e const litEncType = (SymbolEncodingType_e)(istart[0] & 3);
size_t const blockSizeMax = ZSTD_blockSizeMax(dctx);
switch(litEncType)
{
@@ -140,7 +156,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
U32 const lhlCode = (istart[0] >> 2) & 3;
U32 const lhc = MEM_readLE32(istart);
size_t hufSuccess;
size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
int const flags = 0
| (ZSTD_DCtx_get_bmi2(dctx) ? HUF_flags_bmi2 : 0)
| (dctx->disableHufAsm ? HUF_flags_disableAsm : 0);
@@ -167,7 +183,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
break;
}
RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
if (!singleStream)
RETURN_ERROR_IF(litSize < MIN_LITERALS_FOR_4_STREAMS, literals_headerWrong,
"Not enough literals (%zu) for the 4-streams mode (min %u)",
@@ -214,10 +230,12 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
}
if (dctx->litBufferLocation == ZSTD_split)
{
assert(litSize > ZSTD_LITBUFFEREXTRASIZE);
ZSTD_memcpy(dctx->litExtraBuffer, dctx->litBufferEnd - ZSTD_LITBUFFEREXTRASIZE, ZSTD_LITBUFFEREXTRASIZE);
ZSTD_memmove(dctx->litBuffer + ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH, dctx->litBuffer, litSize - ZSTD_LITBUFFEREXTRASIZE);
dctx->litBuffer += ZSTD_LITBUFFEREXTRASIZE - WILDCOPY_OVERLENGTH;
dctx->litBufferEnd -= WILDCOPY_OVERLENGTH;
assert(dctx->litBufferEnd <= (BYTE*)dst + blockSizeMax);
}
RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected, "");
@@ -232,7 +250,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
case set_basic:
{ size_t litSize, lhSize;
U32 const lhlCode = ((istart[0]) >> 2) & 3;
size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
switch(lhlCode)
{
case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */
@@ -251,6 +269,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
}
RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
RETURN_ERROR_IF(expectedWriteSize < litSize, dstSize_tooSmall, "");
ZSTD_allocateLiteralsBuffer(dctx, dst, dstCapacity, litSize, streaming, expectedWriteSize, 1);
if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */
@@ -279,7 +298,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
case set_rle:
{ U32 const lhlCode = ((istart[0]) >> 2) & 3;
size_t litSize, lhSize;
size_t expectedWriteSize = MIN(ZSTD_BLOCKSIZE_MAX, dstCapacity);
size_t expectedWriteSize = MIN(blockSizeMax, dstCapacity);
switch(lhlCode)
{
case 0: case 2: default: /* note : default is impossible, since lhlCode into [0..3] */
@@ -298,7 +317,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
break;
}
RETURN_ERROR_IF(litSize > 0 && dst == NULL, dstSize_tooSmall, "NULL not handled");
RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected, "");
RETURN_ERROR_IF(litSize > blockSizeMax, corruption_detected, "");
RETURN_ERROR_IF(expectedWriteSize < litSize, dstSize_tooSmall, "");
ZSTD_allocateLiteralsBuffer(dctx, dst, dstCapacity, litSize, streaming, expectedWriteSize, 1);
if (dctx->litBufferLocation == ZSTD_split)
@@ -320,6 +339,18 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
}
}
/* Hidden declaration for fullbench */
size_t ZSTD_decodeLiteralsBlock_wrapper(ZSTD_DCtx* dctx,
const void* src, size_t srcSize,
void* dst, size_t dstCapacity);
size_t ZSTD_decodeLiteralsBlock_wrapper(ZSTD_DCtx* dctx,
const void* src, size_t srcSize,
void* dst, size_t dstCapacity)
{
dctx->isFrameDecompression = 0;
return ZSTD_decodeLiteralsBlock(dctx, src, srcSize, dst, dstCapacity, not_streaming);
}
/* Default FSE distribution tables.
* These are pre-calculated FSE decoding tables using default distributions as defined in specification :
* https://github.com/facebook/zstd/blob/release/doc/zstd_compression_format.md#default-distributions
@@ -327,7 +358,7 @@ size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
* - start from default distributions, present in /lib/common/zstd_internal.h
* - generate tables normally, using ZSTD_buildFSETable()
* - printout the content of tables
* - pretify output, report below, test with fuzzer to ensure it's correct */
* - prettify output, report below, test with fuzzer to ensure it's correct */
/* Default FSE distribution table for Literal Lengths */
static const ZSTD_seqSymbol LL_defaultDTable[(1<<LL_DEFAULTNORMLOG)+1] = {
@@ -614,7 +645,7 @@ void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,
* @return : nb bytes read from src,
* or an error code if it fails */
static size_t ZSTD_buildSeqTable(ZSTD_seqSymbol* DTableSpace, const ZSTD_seqSymbol** DTablePtr,
symbolEncodingType_e type, unsigned max, U32 maxLog,
SymbolEncodingType_e type, unsigned max, U32 maxLog,
const void* src, size_t srcSize,
const U32* baseValue, const U8* nbAdditionalBits,
const ZSTD_seqSymbol* defaultTable, U32 flagRepeatTable,
@@ -675,11 +706,6 @@ size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
/* SeqHead */
nbSeq = *ip++;
if (!nbSeq) {
*nbSeqPtr=0;
RETURN_ERROR_IF(srcSize != 1, srcSize_wrong, "");
return 1;
}
if (nbSeq > 0x7F) {
if (nbSeq == 0xFF) {
RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong, "");
@@ -692,11 +718,19 @@ size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
}
*nbSeqPtr = nbSeq;
if (nbSeq == 0) {
/* No sequence : section ends immediately */
RETURN_ERROR_IF(ip != iend, corruption_detected,
"extraneous data present in the Sequences section");
return (size_t)(ip - istart);
}
/* FSE table descriptors */
RETURN_ERROR_IF(ip+1 > iend, srcSize_wrong, ""); /* minimum possible size: 1 byte for symbol encoding types */
{ symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
RETURN_ERROR_IF(*ip & 3, corruption_detected, ""); /* The last field, Reserved, must be all-zeroes. */
{ SymbolEncodingType_e const LLtype = (SymbolEncodingType_e)(*ip >> 6);
SymbolEncodingType_e const OFtype = (SymbolEncodingType_e)((*ip >> 4) & 3);
SymbolEncodingType_e const MLtype = (SymbolEncodingType_e)((*ip >> 2) & 3);
ip++;
/* Build DTables */
@@ -840,7 +874,7 @@ static void ZSTD_safecopy(BYTE* op, const BYTE* const oend_w, BYTE const* ip, pt
/* ZSTD_safecopyDstBeforeSrc():
* This version allows overlap with dst before src, or handles the non-overlap case with dst after src
* Kept separate from more common ZSTD_safecopy case to avoid performance impact to the safecopy common case */
static void ZSTD_safecopyDstBeforeSrc(BYTE* op, BYTE const* ip, ptrdiff_t length) {
static void ZSTD_safecopyDstBeforeSrc(BYTE* op, const BYTE* ip, ptrdiff_t length) {
ptrdiff_t const diff = op - ip;
BYTE* const oend = op + length;
@@ -869,6 +903,7 @@ static void ZSTD_safecopyDstBeforeSrc(BYTE* op, BYTE const* ip, ptrdiff_t length
* to be optimized for many small sequences, since those fall into ZSTD_execSequence().
*/
FORCE_NOINLINE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_execSequenceEnd(BYTE* op,
BYTE* const oend, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit,
@@ -916,6 +951,7 @@ size_t ZSTD_execSequenceEnd(BYTE* op,
* This version is intended to be used during instances where the litBuffer is still split. It is kept separate to avoid performance impact for the good case.
*/
FORCE_NOINLINE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_execSequenceEndSplitLitBuffer(BYTE* op,
BYTE* const oend, const BYTE* const oend_w, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit,
@@ -961,6 +997,7 @@ size_t ZSTD_execSequenceEndSplitLitBuffer(BYTE* op,
}
HINT_INLINE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_execSequence(BYTE* op,
BYTE* const oend, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit,
@@ -1059,6 +1096,7 @@ size_t ZSTD_execSequence(BYTE* op,
}
HINT_INLINE
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
size_t ZSTD_execSequenceSplitLitBuffer(BYTE* op,
BYTE* const oend, const BYTE* const oend_w, seq_t sequence,
const BYTE** litPtr, const BYTE* const litLimit,
@@ -1181,14 +1219,20 @@ ZSTD_updateFseStateWithDInfo(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, U16
typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
/**
* ZSTD_decodeSequence():
* @p longOffsets : tells the decoder to reload more bit while decoding large offsets
* only used in 32-bit mode
* @return : Sequence (litL + matchL + offset)
*/
FORCE_INLINE_TEMPLATE seq_t
ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets, const int isLastSeq)
{
seq_t seq;
/*
* ZSTD_seqSymbol is a structure with a total of 64 bits wide. So it can be
* loaded in one operation and extracted its fields by simply shifting or
* bit-extracting on aarch64.
* ZSTD_seqSymbol is a 64 bits wide structure.
* It can be loaded in one operation
* and its fields extracted by simply shifting or bit-extracting on aarch64.
* GCC doesn't recognize this and generates more unnecessary ldr/ldrb/ldrh
* operations that cause performance drop. This can be avoided by using this
* ZSTD_memcpy hack.
@@ -1261,7 +1305,7 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
} else {
offset = ofBase + ll0 + BIT_readBitsFast(&seqState->DStream, 1);
{ size_t temp = (offset==3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */
temp -= !temp; /* 0 is not valid: input corrupted => force offset to -1 => corruption detected at execSequence */
if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
seqState->prevOffset[1] = seqState->prevOffset[0];
seqState->prevOffset[0] = offset = temp;
@@ -1288,17 +1332,22 @@ ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
(U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
ZSTD_updateFseStateWithDInfo(&seqState->stateLL, &seqState->DStream, llNext, llnbBits); /* <= 9 bits */
ZSTD_updateFseStateWithDInfo(&seqState->stateML, &seqState->DStream, mlNext, mlnbBits); /* <= 9 bits */
if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
ZSTD_updateFseStateWithDInfo(&seqState->stateOffb, &seqState->DStream, ofNext, ofnbBits); /* <= 8 bits */
if (!isLastSeq) {
/* don't update FSE state for last Sequence */
ZSTD_updateFseStateWithDInfo(&seqState->stateLL, &seqState->DStream, llNext, llnbBits); /* <= 9 bits */
ZSTD_updateFseStateWithDInfo(&seqState->stateML, &seqState->DStream, mlNext, mlnbBits); /* <= 9 bits */
if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
ZSTD_updateFseStateWithDInfo(&seqState->stateOffb, &seqState->DStream, ofNext, ofnbBits); /* <= 8 bits */
BIT_reloadDStream(&seqState->DStream);
}
}
return seq;
}
#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
MEM_STATIC int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefixStart, BYTE const* oLitEnd)
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
#if DEBUGLEVEL >= 1
static int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefixStart, BYTE const* oLitEnd)
{
size_t const windowSize = dctx->fParams.windowSize;
/* No dictionary used. */
@@ -1312,30 +1361,33 @@ MEM_STATIC int ZSTD_dictionaryIsActive(ZSTD_DCtx const* dctx, BYTE const* prefix
/* Dictionary is active. */
return 1;
}
#endif
MEM_STATIC void ZSTD_assertValidSequence(
static void ZSTD_assertValidSequence(
ZSTD_DCtx const* dctx,
BYTE const* op, BYTE const* oend,
seq_t const seq,
BYTE const* prefixStart, BYTE const* virtualStart)
{
#if DEBUGLEVEL >= 1
size_t const windowSize = dctx->fParams.windowSize;
size_t const sequenceSize = seq.litLength + seq.matchLength;
BYTE const* const oLitEnd = op + seq.litLength;
DEBUGLOG(6, "Checking sequence: litL=%u matchL=%u offset=%u",
(U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
assert(op <= oend);
assert((size_t)(oend - op) >= sequenceSize);
assert(sequenceSize <= ZSTD_BLOCKSIZE_MAX);
if (ZSTD_dictionaryIsActive(dctx, prefixStart, oLitEnd)) {
size_t const dictSize = (size_t)((char const*)dctx->dictContentEndForFuzzing - (char const*)dctx->dictContentBeginForFuzzing);
/* Offset must be within the dictionary. */
assert(seq.offset <= (size_t)(oLitEnd - virtualStart));
assert(seq.offset <= windowSize + dictSize);
} else {
/* Offset must be within our window. */
assert(seq.offset <= windowSize);
if (dctx->isFrameDecompression) {
size_t const windowSize = dctx->fParams.windowSize;
size_t const sequenceSize = seq.litLength + seq.matchLength;
BYTE const* const oLitEnd = op + seq.litLength;
DEBUGLOG(6, "Checking sequence: litL=%u matchL=%u offset=%u",
(U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
assert(op <= oend);
assert((size_t)(oend - op) >= sequenceSize);
assert(sequenceSize <= ZSTD_blockSizeMax(dctx));
if (ZSTD_dictionaryIsActive(dctx, prefixStart, oLitEnd)) {
size_t const dictSize = (size_t)((char const*)dctx->dictContentEndForFuzzing - (char const*)dctx->dictContentBeginForFuzzing);
/* Offset must be within the dictionary. */
assert(seq.offset <= (size_t)(oLitEnd - virtualStart));
assert(seq.offset <= windowSize + dictSize);
} else {
/* Offset must be within our window. */
assert(seq.offset <= windowSize);
}
}
#else
(void)dctx, (void)op, (void)oend, (void)seq, (void)prefixStart, (void)virtualStart;
@@ -1351,23 +1403,21 @@ DONT_VECTORIZE
ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
const BYTE* ip = (const BYTE*)seqStart;
const BYTE* const iend = ip + seqSize;
BYTE* const ostart = (BYTE*)dst;
BYTE* const oend = ostart + maxDstSize;
BYTE* const oend = ZSTD_maybeNullPtrAdd(ostart, maxDstSize);
BYTE* op = ostart;
const BYTE* litPtr = dctx->litPtr;
const BYTE* litBufferEnd = dctx->litBufferEnd;
const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
const BYTE* const vBase = (const BYTE*) (dctx->virtualStart);
const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer");
(void)frame;
DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer (%i seqs)", nbSeq);
/* Regen sequences */
/* Literals are split between internal buffer & output buffer */
if (nbSeq) {
seqState_t seqState;
dctx->fseEntropy = 1;
@@ -1386,8 +1436,7 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
BIT_DStream_completed < BIT_DStream_overflow);
/* decompress without overrunning litPtr begins */
{
seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
{ seq_t sequence = {0,0,0}; /* some static analyzer believe that @sequence is not initialized (it necessarily is, since for(;;) loop as at least one iteration) */
/* Align the decompression loop to 32 + 16 bytes.
*
* zstd compiled with gcc-9 on an Intel i9-9900k shows 10% decompression
@@ -1449,27 +1498,26 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
#endif
/* Handle the initial state where litBuffer is currently split between dst and litExtraBuffer */
for (; litPtr + sequence.litLength <= dctx->litBufferEnd; ) {
size_t const oneSeqSize = ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequence.litLength - WILDCOPY_OVERLENGTH, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
for ( ; nbSeq; nbSeq--) {
sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
if (litPtr + sequence.litLength > dctx->litBufferEnd) break;
{ size_t const oneSeqSize = ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequence.litLength - WILDCOPY_OVERLENGTH, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
assert(!ZSTD_isError(oneSeqSize));
ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
#endif
if (UNLIKELY(ZSTD_isError(oneSeqSize)))
return oneSeqSize;
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
op += oneSeqSize;
if (UNLIKELY(!--nbSeq))
break;
BIT_reloadDStream(&(seqState.DStream));
sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
}
if (UNLIKELY(ZSTD_isError(oneSeqSize)))
return oneSeqSize;
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
op += oneSeqSize;
} }
DEBUGLOG(6, "reached: (litPtr + sequence.litLength > dctx->litBufferEnd)");
/* If there are more sequences, they will need to read literals from litExtraBuffer; copy over the remainder from dst and update litPtr and litEnd */
if (nbSeq > 0) {
const size_t leftoverLit = dctx->litBufferEnd - litPtr;
if (leftoverLit)
{
DEBUGLOG(6, "There are %i sequences left, and %zu/%zu literals left in buffer", nbSeq, leftoverLit, sequence.litLength);
if (leftoverLit) {
RETURN_ERROR_IF(leftoverLit > (size_t)(oend - op), dstSize_tooSmall, "remaining lit must fit within dstBuffer");
ZSTD_safecopyDstBeforeSrc(op, litPtr, leftoverLit);
sequence.litLength -= leftoverLit;
@@ -1478,24 +1526,22 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
litPtr = dctx->litExtraBuffer;
litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
dctx->litBufferLocation = ZSTD_not_in_dst;
{
size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
{ size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
#endif
if (UNLIKELY(ZSTD_isError(oneSeqSize)))
return oneSeqSize;
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
op += oneSeqSize;
if (--nbSeq)
BIT_reloadDStream(&(seqState.DStream));
}
nbSeq--;
}
}
if (nbSeq > 0) /* there is remaining lit from extra buffer */
{
if (nbSeq > 0) {
/* there is remaining lit from extra buffer */
#if defined(__GNUC__) && defined(__x86_64__)
__asm__(".p2align 6");
@@ -1514,35 +1560,34 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
# endif
#endif
for (; ; ) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
for ( ; nbSeq ; nbSeq--) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litBufferEnd, prefixStart, vBase, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
#endif
if (UNLIKELY(ZSTD_isError(oneSeqSize)))
return oneSeqSize;
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
op += oneSeqSize;
if (UNLIKELY(!--nbSeq))
break;
BIT_reloadDStream(&(seqState.DStream));
}
}
/* check if reached exact end */
DEBUGLOG(5, "ZSTD_decompressSequences_bodySplitLitBuffer: after decode loop, remaining nbSeq : %i", nbSeq);
RETURN_ERROR_IF(nbSeq, corruption_detected, "");
RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected, "");
DEBUGLOG(5, "bitStream : start=%p, ptr=%p, bitsConsumed=%u", seqState.DStream.start, seqState.DStream.ptr, seqState.DStream.bitsConsumed);
RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
/* save reps for next block */
{ U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
}
/* last literal segment */
if (dctx->litBufferLocation == ZSTD_split) /* split hasn't been reached yet, first get dst then copy litExtraBuffer */
{
size_t const lastLLSize = litBufferEnd - litPtr;
if (dctx->litBufferLocation == ZSTD_split) {
/* split hasn't been reached yet, first get dst then copy litExtraBuffer */
size_t const lastLLSize = (size_t)(litBufferEnd - litPtr);
DEBUGLOG(6, "copy last literals from segment : %u", (U32)lastLLSize);
RETURN_ERROR_IF(lastLLSize > (size_t)(oend - op), dstSize_tooSmall, "");
if (op != NULL) {
ZSTD_memmove(op, litPtr, lastLLSize);
@@ -1552,15 +1597,17 @@ ZSTD_decompressSequences_bodySplitLitBuffer( ZSTD_DCtx* dctx,
litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
dctx->litBufferLocation = ZSTD_not_in_dst;
}
{ size_t const lastLLSize = litBufferEnd - litPtr;
/* copy last literals from internal buffer */
{ size_t const lastLLSize = (size_t)(litBufferEnd - litPtr);
DEBUGLOG(6, "copy last literals from internal buffer : %u", (U32)lastLLSize);
RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
if (op != NULL) {
ZSTD_memcpy(op, litPtr, lastLLSize);
op += lastLLSize;
}
}
} }
return op-ostart;
DEBUGLOG(6, "decoded block of size %u bytes", (U32)(op - ostart));
return (size_t)(op - ostart);
}
FORCE_INLINE_TEMPLATE size_t
@@ -1568,13 +1615,12 @@ DONT_VECTORIZE
ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
const BYTE* ip = (const BYTE*)seqStart;
const BYTE* const iend = ip + seqSize;
BYTE* const ostart = (BYTE*)dst;
BYTE* const oend = dctx->litBufferLocation == ZSTD_not_in_dst ? ostart + maxDstSize : dctx->litBuffer;
BYTE* const oend = dctx->litBufferLocation == ZSTD_not_in_dst ? ZSTD_maybeNullPtrAdd(ostart, maxDstSize) : dctx->litBuffer;
BYTE* op = ostart;
const BYTE* litPtr = dctx->litPtr;
const BYTE* const litEnd = litPtr + dctx->litSize;
@@ -1582,7 +1628,6 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
const BYTE* const vBase = (const BYTE*)(dctx->virtualStart);
const BYTE* const dictEnd = (const BYTE*)(dctx->dictEnd);
DEBUGLOG(5, "ZSTD_decompressSequences_body: nbSeq = %d", nbSeq);
(void)frame;
/* Regen sequences */
if (nbSeq) {
@@ -1597,11 +1642,6 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
assert(dst != NULL);
ZSTD_STATIC_ASSERT(
BIT_DStream_unfinished < BIT_DStream_completed &&
BIT_DStream_endOfBuffer < BIT_DStream_completed &&
BIT_DStream_completed < BIT_DStream_overflow);
#if defined(__GNUC__) && defined(__x86_64__)
__asm__(".p2align 6");
__asm__("nop");
@@ -1616,73 +1656,70 @@ ZSTD_decompressSequences_body(ZSTD_DCtx* dctx,
# endif
#endif
for ( ; ; ) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
for ( ; nbSeq ; nbSeq--) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, nbSeq==1);
size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, prefixStart, vBase, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
ZSTD_assertValidSequence(dctx, op, oend, sequence, prefixStart, vBase);
#endif
if (UNLIKELY(ZSTD_isError(oneSeqSize)))
return oneSeqSize;
DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
op += oneSeqSize;
if (UNLIKELY(!--nbSeq))
break;
BIT_reloadDStream(&(seqState.DStream));
}
/* check if reached exact end */
DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);
RETURN_ERROR_IF(nbSeq, corruption_detected, "");
RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected, "");
assert(nbSeq == 0);
RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
/* save reps for next block */
{ U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
}
/* last literal segment */
{ size_t const lastLLSize = litEnd - litPtr;
{ size_t const lastLLSize = (size_t)(litEnd - litPtr);
DEBUGLOG(6, "copy last literals : %u", (U32)lastLLSize);
RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall, "");
if (op != NULL) {
ZSTD_memcpy(op, litPtr, lastLLSize);
op += lastLLSize;
}
}
} }
return op-ostart;
DEBUGLOG(6, "decoded block of size %u bytes", (U32)(op - ostart));
return (size_t)(op - ostart);
}
static size_t
ZSTD_decompressSequences_default(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
static size_t
ZSTD_decompressSequencesSplitLitBuffer_default(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
FORCE_INLINE_TEMPLATE size_t
ZSTD_prefetchMatch(size_t prefetchPos, seq_t const sequence,
FORCE_INLINE_TEMPLATE
size_t ZSTD_prefetchMatch(size_t prefetchPos, seq_t const sequence,
const BYTE* const prefixStart, const BYTE* const dictEnd)
{
prefetchPos += sequence.litLength;
{ const BYTE* const matchBase = (sequence.offset > prefetchPos) ? dictEnd : prefixStart;
const BYTE* const match = matchBase + prefetchPos - sequence.offset; /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
* No consequence though : memory address is only used for prefetching, not for dereferencing */
/* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
* No consequence though : memory address is only used for prefetching, not for dereferencing */
const BYTE* const match = ZSTD_wrappedPtrSub(ZSTD_wrappedPtrAdd(matchBase, prefetchPos), sequence.offset);
PREFETCH_L1(match); PREFETCH_L1(match+CACHELINE_SIZE); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
}
return prefetchPos + sequence.matchLength;
@@ -1697,20 +1734,18 @@ ZSTD_decompressSequencesLong_body(
ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
const BYTE* ip = (const BYTE*)seqStart;
const BYTE* const iend = ip + seqSize;
BYTE* const ostart = (BYTE*)dst;
BYTE* const oend = dctx->litBufferLocation == ZSTD_in_dst ? dctx->litBuffer : ostart + maxDstSize;
BYTE* const oend = dctx->litBufferLocation == ZSTD_in_dst ? dctx->litBuffer : ZSTD_maybeNullPtrAdd(ostart, maxDstSize);
BYTE* op = ostart;
const BYTE* litPtr = dctx->litPtr;
const BYTE* litBufferEnd = dctx->litBufferEnd;
const BYTE* const prefixStart = (const BYTE*) (dctx->prefixStart);
const BYTE* const dictStart = (const BYTE*) (dctx->virtualStart);
const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
(void)frame;
/* Regen sequences */
if (nbSeq) {
@@ -1735,20 +1770,17 @@ ZSTD_decompressSequencesLong_body(
ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
/* prepare in advance */
for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
for (seqNb=0; seqNb<seqAdvance; seqNb++) {
seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset, seqNb == nbSeq-1);
prefetchPos = ZSTD_prefetchMatch(prefetchPos, sequence, prefixStart, dictEnd);
sequences[seqNb] = sequence;
}
RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected, "");
/* decompress without stomping litBuffer */
for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb < nbSeq); seqNb++) {
seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
size_t oneSeqSize;
for (; seqNb < nbSeq; seqNb++) {
seq_t sequence = ZSTD_decodeSequence(&seqState, isLongOffset, seqNb == nbSeq-1);
if (dctx->litBufferLocation == ZSTD_split && litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength > dctx->litBufferEnd)
{
if (dctx->litBufferLocation == ZSTD_split && litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength > dctx->litBufferEnd) {
/* lit buffer is reaching split point, empty out the first buffer and transition to litExtraBuffer */
const size_t leftoverLit = dctx->litBufferEnd - litPtr;
if (leftoverLit)
@@ -1761,26 +1793,26 @@ ZSTD_decompressSequencesLong_body(
litPtr = dctx->litExtraBuffer;
litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
dctx->litBufferLocation = ZSTD_not_in_dst;
oneSeqSize = ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
{ size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
assert(!ZSTD_isError(oneSeqSize));
ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
#endif
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
prefetchPos = ZSTD_prefetchMatch(prefetchPos, sequence, prefixStart, dictEnd);
sequences[seqNb & STORED_SEQS_MASK] = sequence;
op += oneSeqSize;
}
prefetchPos = ZSTD_prefetchMatch(prefetchPos, sequence, prefixStart, dictEnd);
sequences[seqNb & STORED_SEQS_MASK] = sequence;
op += oneSeqSize;
} }
else
{
/* lit buffer is either wholly contained in first or second split, or not split at all*/
oneSeqSize = dctx->litBufferLocation == ZSTD_split ?
size_t const oneSeqSize = dctx->litBufferLocation == ZSTD_split ?
ZSTD_execSequenceSplitLitBuffer(op, oend, litPtr + sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK].litLength - WILDCOPY_OVERLENGTH, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd) :
ZSTD_execSequence(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
ZSTD_assertValidSequence(dctx, op, oend, sequences[(seqNb - ADVANCED_SEQS) & STORED_SEQS_MASK], prefixStart, dictStart);
#endif
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
@@ -1789,17 +1821,15 @@ ZSTD_decompressSequencesLong_body(
op += oneSeqSize;
}
}
RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected, "");
RETURN_ERROR_IF(!BIT_endOfDStream(&seqState.DStream), corruption_detected, "");
/* finish queue */
seqNb -= seqAdvance;
for ( ; seqNb<nbSeq ; seqNb++) {
seq_t *sequence = &(sequences[seqNb&STORED_SEQS_MASK]);
if (dctx->litBufferLocation == ZSTD_split && litPtr + sequence->litLength > dctx->litBufferEnd)
{
if (dctx->litBufferLocation == ZSTD_split && litPtr + sequence->litLength > dctx->litBufferEnd) {
const size_t leftoverLit = dctx->litBufferEnd - litPtr;
if (leftoverLit)
{
if (leftoverLit) {
RETURN_ERROR_IF(leftoverLit > (size_t)(oend - op), dstSize_tooSmall, "remaining lit must fit within dstBuffer");
ZSTD_safecopyDstBeforeSrc(op, litPtr, leftoverLit);
sequence->litLength -= leftoverLit;
@@ -1808,11 +1838,10 @@ ZSTD_decompressSequencesLong_body(
litPtr = dctx->litExtraBuffer;
litBufferEnd = dctx->litExtraBuffer + ZSTD_LITBUFFEREXTRASIZE;
dctx->litBufferLocation = ZSTD_not_in_dst;
{
size_t const oneSeqSize = ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
{ size_t const oneSeqSize = ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
#endif
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
op += oneSeqSize;
@@ -1825,7 +1854,7 @@ ZSTD_decompressSequencesLong_body(
ZSTD_execSequence(op, oend, *sequence, &litPtr, litBufferEnd, prefixStart, dictStart, dictEnd);
#if defined(FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION) && defined(FUZZING_ASSERT_VALID_SEQUENCE)
assert(!ZSTD_isError(oneSeqSize));
if (frame) ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
ZSTD_assertValidSequence(dctx, op, oend, sequences[seqNb&STORED_SEQS_MASK], prefixStart, dictStart);
#endif
if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
op += oneSeqSize;
@@ -1837,8 +1866,7 @@ ZSTD_decompressSequencesLong_body(
}
/* last literal segment */
if (dctx->litBufferLocation == ZSTD_split) /* first deplete literal buffer in dst, then copy litExtraBuffer */
{
if (dctx->litBufferLocation == ZSTD_split) { /* first deplete literal buffer in dst, then copy litExtraBuffer */
size_t const lastLLSize = litBufferEnd - litPtr;
RETURN_ERROR_IF(lastLLSize > (size_t)(oend - op), dstSize_tooSmall, "");
if (op != NULL) {
@@ -1856,17 +1884,16 @@ ZSTD_decompressSequencesLong_body(
}
}
return op-ostart;
return (size_t)(op - ostart);
}
static size_t
ZSTD_decompressSequencesLong_default(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
@@ -1880,20 +1907,18 @@ DONT_VECTORIZE
ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
static BMI2_TARGET_ATTRIBUTE size_t
DONT_VECTORIZE
ZSTD_decompressSequencesSplitLitBuffer_bmi2(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_bodySplitLitBuffer(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
@@ -1902,50 +1927,40 @@ static BMI2_TARGET_ATTRIBUTE size_t
ZSTD_decompressSequencesLong_bmi2(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesLong_body(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
#endif /* DYNAMIC_BMI2 */
typedef size_t (*ZSTD_decompressSequences_t)(
ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame);
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
static size_t
ZSTD_decompressSequences(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
DEBUGLOG(5, "ZSTD_decompressSequences");
#if DYNAMIC_BMI2
if (ZSTD_DCtx_get_bmi2(dctx)) {
return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif
return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
static size_t
ZSTD_decompressSequencesSplitLitBuffer(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
DEBUGLOG(5, "ZSTD_decompressSequencesSplitLitBuffer");
#if DYNAMIC_BMI2
if (ZSTD_DCtx_get_bmi2(dctx)) {
return ZSTD_decompressSequencesSplitLitBuffer_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesSplitLitBuffer_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif
return ZSTD_decompressSequencesSplitLitBuffer_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesSplitLitBuffer_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG */
@@ -1960,16 +1975,15 @@ static size_t
ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
void* dst, size_t maxDstSize,
const void* seqStart, size_t seqSize, int nbSeq,
const ZSTD_longOffset_e isLongOffset,
const int frame)
const ZSTD_longOffset_e isLongOffset)
{
DEBUGLOG(5, "ZSTD_decompressSequencesLong");
#if DYNAMIC_BMI2
if (ZSTD_DCtx_get_bmi2(dctx)) {
return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif
return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, nbSeq, isLongOffset);
}
#endif /* ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT */
@@ -2051,20 +2065,20 @@ static size_t ZSTD_maxShortOffset(void)
size_t
ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const int frame, const streaming_operation streaming)
const void* src, size_t srcSize, const streaming_operation streaming)
{ /* blockType == blockCompressed */
const BYTE* ip = (const BYTE*)src;
DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
DEBUGLOG(5, "ZSTD_decompressBlock_internal (cSize : %u)", (unsigned)srcSize);
/* Note : the wording of the specification
* allows compressed block to be sized exactly ZSTD_BLOCKSIZE_MAX.
* allows compressed block to be sized exactly ZSTD_blockSizeMax(dctx).
* This generally does not happen, as it makes little sense,
* since an uncompressed block would feature same size and have no decompression cost.
* Also, note that decoder from reference libzstd before < v1.5.4
* would consider this edge case as an error.
* As a consequence, avoid generating compressed blocks of size ZSTD_BLOCKSIZE_MAX
* As a consequence, avoid generating compressed blocks of size ZSTD_blockSizeMax(dctx)
* for broader compatibility with the deployed ecosystem of zstd decoders */
RETURN_ERROR_IF(srcSize > ZSTD_BLOCKSIZE_MAX, srcSize_wrong, "");
RETURN_ERROR_IF(srcSize > ZSTD_blockSizeMax(dctx), srcSize_wrong, "");
/* Decode literals section */
{ size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize, dst, dstCapacity, streaming);
@@ -2079,8 +2093,8 @@ ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
/* Compute the maximum block size, which must also work when !frame and fParams are unset.
* Additionally, take the min with dstCapacity to ensure that the totalHistorySize fits in a size_t.
*/
size_t const blockSizeMax = MIN(dstCapacity, (frame ? dctx->fParams.blockSizeMax : ZSTD_BLOCKSIZE_MAX));
size_t const totalHistorySize = ZSTD_totalHistorySize((BYTE*)dst + blockSizeMax, (BYTE const*)dctx->virtualStart);
size_t const blockSizeMax = MIN(dstCapacity, ZSTD_blockSizeMax(dctx));
size_t const totalHistorySize = ZSTD_totalHistorySize(ZSTD_maybeNullPtrAdd((BYTE*)dst, blockSizeMax), (BYTE const*)dctx->virtualStart);
/* isLongOffset must be true if there are long offsets.
* Offsets are long if they are larger than ZSTD_maxShortOffset().
* We don't expect that to be the case in 64-bit mode.
@@ -2145,21 +2159,22 @@ ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
{
#endif
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
#endif
}
#ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
/* else */
if (dctx->litBufferLocation == ZSTD_split)
return ZSTD_decompressSequencesSplitLitBuffer(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequencesSplitLitBuffer(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
else
return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset, frame);
return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize, nbSeq, isLongOffset);
#endif
}
}
ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize)
{
if (dst != dctx->previousDstEnd && dstSize > 0) { /* not contiguous */
@@ -2176,8 +2191,10 @@ size_t ZSTD_decompressBlock_deprecated(ZSTD_DCtx* dctx,
const void* src, size_t srcSize)
{
size_t dSize;
dctx->isFrameDecompression = 0;
ZSTD_checkContinuity(dctx, dst, dstCapacity);
dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, /* frame */ 0, not_streaming);
dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize, not_streaming);
FORWARD_IF_ERROR(dSize, "");
dctx->previousDstEnd = (char*)dst + dSize;
return dSize;
}
Utilities/cmzstd/lib/decompress/zstd_decompress_block.h
+1 -1
View File
@@ -47,7 +47,7 @@ typedef enum {
*/
size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize, const int frame, const streaming_operation streaming);
const void* src, size_t srcSize, const streaming_operation streaming);
/* ZSTD_buildFSETable() :
* generate FSE decoding table for one symbol (ll, ml or off)
Utilities/cmzstd/lib/decompress/zstd_decompress_internal.h
+7 -5
View File
@@ -136,7 +136,7 @@ struct ZSTD_DCtx_s
const void* virtualStart; /* virtual start of previous segment if it was just before current one */
const void* dictEnd; /* end of previous segment */
size_t expected;
ZSTD_frameHeader fParams;
ZSTD_FrameHeader fParams;
U64 processedCSize;
U64 decodedSize;
blockType_e bType; /* used in ZSTD_decompressContinue(), store blockType between block header decoding and block decompression stages */
@@ -153,7 +153,8 @@ struct ZSTD_DCtx_s
size_t litSize;
size_t rleSize;
size_t staticSize;
#if DYNAMIC_BMI2 != 0
int isFrameDecompression;
#if DYNAMIC_BMI2
int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
#endif
@@ -166,6 +167,7 @@ struct ZSTD_DCtx_s
ZSTD_DDictHashSet* ddictSet; /* Hash set for multiple ddicts */
ZSTD_refMultipleDDicts_e refMultipleDDicts; /* User specified: if == 1, will allow references to multiple DDicts. Default == 0 (disabled) */
int disableHufAsm;
int maxBlockSizeParam;
/* streaming */
ZSTD_dStreamStage streamStage;
@@ -209,11 +211,11 @@ struct ZSTD_DCtx_s
}; /* typedef'd to ZSTD_DCtx within "zstd.h" */
MEM_STATIC int ZSTD_DCtx_get_bmi2(const struct ZSTD_DCtx_s *dctx) {
#if DYNAMIC_BMI2 != 0
return dctx->bmi2;
#if DYNAMIC_BMI2
return dctx->bmi2;
#else
(void)dctx;
return 0;
return 0;
#endif
}
Utilities/cmzstd/lib/dictBuilder/cover.c
+76 -31
View File
@@ -21,8 +21,17 @@
/*-*************************************
* Dependencies
***************************************/
/* qsort_r is an extension. */
#if defined(__linux) || defined(__linux__) || defined(linux) || defined(__gnu_linux__) || \
defined(__CYGWIN__) || defined(__MSYS__)
#if !defined(_GNU_SOURCE) && !defined(__ANDROID__) /* NDK doesn't ship qsort_r(). */
#define _GNU_SOURCE
#endif
#endif
#include <stdio.h> /* fprintf */
#include <stdlib.h> /* malloc, free, qsort */
#include <stdlib.h> /* malloc, free, qsort_r */
#include <string.h> /* memset */
#include <time.h> /* clock */
@@ -31,8 +40,8 @@
#endif
#include "../common/mem.h" /* read */
#include "../common/pool.h"
#include "../common/threading.h"
#include "../common/pool.h" /* POOL_ctx */
#include "../common/threading.h" /* ZSTD_pthread_mutex_t */
#include "../common/zstd_internal.h" /* includes zstd.h */
#include "../common/bits.h" /* ZSTD_highbit32 */
#include "../zdict.h"
@@ -78,7 +87,7 @@ static clock_t g_time = 0;
#undef LOCALDISPLAYUPDATE
#define LOCALDISPLAYUPDATE(displayLevel, l, ...) \
if (displayLevel >= l) { \
if ((clock() - g_time > g_refreshRate) || (displayLevel >= 4)) { \
if ((clock() - g_time > g_refreshRate) || (displayLevel >= 4)) { \
g_time = clock(); \
DISPLAY(__VA_ARGS__); \
} \
@@ -232,8 +241,10 @@ typedef struct {
unsigned d;
} COVER_ctx_t;
/* We need a global context for qsort... */
#if !defined(_GNU_SOURCE) && !defined(__APPLE__) && !defined(_MSC_VER)
/* C90 only offers qsort() that needs a global context. */
static COVER_ctx_t *g_coverCtx = NULL;
#endif
/*-*************************************
* Helper functions
@@ -276,11 +287,15 @@ static int COVER_cmp8(COVER_ctx_t *ctx, const void *lp, const void *rp) {
/**
* Same as COVER_cmp() except ties are broken by pointer value
* NOTE: g_coverCtx must be set to call this function. A global is required because
* qsort doesn't take an opaque pointer.
*/
static int WIN_CDECL COVER_strict_cmp(const void *lp, const void *rp) {
int result = COVER_cmp(g_coverCtx, lp, rp);
#if (defined(_WIN32) && defined(_MSC_VER)) || defined(__APPLE__)
static int WIN_CDECL COVER_strict_cmp(void* g_coverCtx, const void* lp, const void* rp) {
#elif defined(_GNU_SOURCE)
static int COVER_strict_cmp(const void *lp, const void *rp, void *g_coverCtx) {
#else /* C90 fallback.*/
static int COVER_strict_cmp(const void *lp, const void *rp) {
#endif
int result = COVER_cmp((COVER_ctx_t*)g_coverCtx, lp, rp);
if (result == 0) {
result = lp < rp ? -1 : 1;
}
@@ -289,21 +304,58 @@ static int WIN_CDECL COVER_strict_cmp(const void *lp, const void *rp) {
/**
* Faster version for d <= 8.
*/
static int WIN_CDECL COVER_strict_cmp8(const void *lp, const void *rp) {
int result = COVER_cmp8(g_coverCtx, lp, rp);
#if (defined(_WIN32) && defined(_MSC_VER)) || defined(__APPLE__)
static int WIN_CDECL COVER_strict_cmp8(void* g_coverCtx, const void* lp, const void* rp) {
#elif defined(_GNU_SOURCE)
static int COVER_strict_cmp8(const void *lp, const void *rp, void *g_coverCtx) {
#else /* C90 fallback.*/
static int COVER_strict_cmp8(const void *lp, const void *rp) {
#endif
int result = COVER_cmp8((COVER_ctx_t*)g_coverCtx, lp, rp);
if (result == 0) {
result = lp < rp ? -1 : 1;
}
return result;
}
/**
* Abstract away divergence of qsort_r() parameters.
* Hopefully when C11 become the norm, we will be able
* to clean it up.
*/
static void stableSort(COVER_ctx_t *ctx) {
#if defined(__APPLE__)
qsort_r(ctx->suffix, ctx->suffixSize, sizeof(U32),
ctx,
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));
#elif defined(_GNU_SOURCE)
qsort_r(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp),
ctx);
#elif defined(_WIN32) && defined(_MSC_VER)
qsort_s(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp),
ctx);
#elif defined(__OpenBSD__)
g_coverCtx = ctx;
mergesort(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));
#else /* C90 fallback.*/
g_coverCtx = ctx;
/* TODO(cavalcanti): implement a reentrant qsort() when is not available. */
qsort(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));
#endif
}
/**
* Returns the first pointer in [first, last) whose element does not compare
* less than value. If no such element exists it returns last.
*/
static const size_t *COVER_lower_bound(const size_t *first, const size_t *last,
static const size_t *COVER_lower_bound(const size_t* first, const size_t* last,
size_t value) {
size_t count = last - first;
size_t count = (size_t)(last - first);
assert(last >= first);
while (count != 0) {
size_t step = count / 2;
const size_t *ptr = first;
@@ -549,7 +601,8 @@ static void COVER_ctx_destroy(COVER_ctx_t *ctx) {
*/
static size_t COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
const size_t *samplesSizes, unsigned nbSamples,
unsigned d, double splitPoint) {
unsigned d, double splitPoint)
{
const BYTE *const samples = (const BYTE *)samplesBuffer;
const size_t totalSamplesSize = COVER_sum(samplesSizes, nbSamples);
/* Split samples into testing and training sets */
@@ -618,17 +671,7 @@ static size_t COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
for (i = 0; i < ctx->suffixSize; ++i) {
ctx->suffix[i] = i;
}
/* qsort doesn't take an opaque pointer, so pass as a global.
* On OpenBSD qsort() is not guaranteed to be stable, their mergesort() is.
*/
g_coverCtx = ctx;
#if defined(__OpenBSD__)
mergesort(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));
#else
qsort(ctx->suffix, ctx->suffixSize, sizeof(U32),
(ctx->d <= 8 ? &COVER_strict_cmp8 : &COVER_strict_cmp));
#endif
stableSort(ctx);
}
DISPLAYLEVEL(2, "Computing frequencies\n");
/* For each dmer group (group of positions with the same first d bytes):
@@ -733,7 +776,7 @@ static size_t COVER_buildDictionary(const COVER_ctx_t *ctx, U32 *freqs,
return tail;
}
ZDICTLIB_API size_t ZDICT_trainFromBuffer_cover(
ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_cover(
void *dictBuffer, size_t dictBufferCapacity,
const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
ZDICT_cover_params_t parameters)
@@ -907,8 +950,10 @@ void COVER_best_start(COVER_best_t *best) {
* Decrements liveJobs and signals any waiting threads if liveJobs == 0.
* If this dictionary is the best so far save it and its parameters.
*/
void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
COVER_dictSelection_t selection) {
void COVER_best_finish(COVER_best_t* best,
ZDICT_cover_params_t parameters,
COVER_dictSelection_t selection)
{
void* dict = selection.dictContent;
size_t compressedSize = selection.totalCompressedSize;
size_t dictSize = selection.dictSize;
@@ -980,8 +1025,8 @@ COVER_dictSelection_t COVER_selectDict(BYTE* customDictContent, size_t dictBuffe
size_t largestCompressed = 0;
BYTE* customDictContentEnd = customDictContent + dictContentSize;
BYTE * largestDictbuffer = (BYTE *)malloc(dictBufferCapacity);
BYTE * candidateDictBuffer = (BYTE *)malloc(dictBufferCapacity);
BYTE* largestDictbuffer = (BYTE*)malloc(dictBufferCapacity);
BYTE* candidateDictBuffer = (BYTE*)malloc(dictBufferCapacity);
double regressionTolerance = ((double)params.shrinkDictMaxRegression / 100.0) + 1.00;
if (!largestDictbuffer || !candidateDictBuffer) {
@@ -1119,7 +1164,7 @@ _cleanup:
free(freqs);
}
ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(
ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_cover(
void* dictBuffer, size_t dictBufferCapacity, const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples,
ZDICT_cover_params_t* parameters)
Utilities/cmzstd/lib/dictBuilder/cover.h
+2 -8
View File
@@ -12,14 +12,8 @@
# define ZDICT_STATIC_LINKING_ONLY
#endif
#include <stdio.h> /* fprintf */
#include <stdlib.h> /* malloc, free, qsort */
#include <string.h> /* memset */
#include <time.h> /* clock */
#include "../common/mem.h" /* read */
#include "../common/pool.h"
#include "../common/threading.h"
#include "../common/zstd_internal.h" /* includes zstd.h */
#include "../common/threading.h" /* ZSTD_pthread_mutex_t */
#include "../common/mem.h" /* U32, BYTE */
#include "../zdict.h"
/**
Utilities/cmzstd/lib/dictBuilder/divsufsort.h
-10
View File
@@ -27,11 +27,6 @@
#ifndef _DIVSUFSORT_H
#define _DIVSUFSORT_H 1
#ifdef __cplusplus
extern "C" {
#endif /* __cplusplus */
/*- Prototypes -*/
/**
@@ -59,9 +54,4 @@ divsufsort(const unsigned char *T, int *SA, int n, int openMP);
int
divbwt(const unsigned char *T, unsigned char *U, int *A, int n, unsigned char * num_indexes, int * indexes, int openMP);
#ifdef __cplusplus
} /* extern "C" */
#endif /* __cplusplus */
#endif /* _DIVSUFSORT_H */
Utilities/cmzstd/lib/dictBuilder/fastcover.c
+2 -2
View File
@@ -545,7 +545,7 @@ FASTCOVER_convertToFastCoverParams(ZDICT_cover_params_t coverParams,
}
ZDICTLIB_API size_t
ZDICTLIB_STATIC_API size_t
ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,
const void* samplesBuffer,
const size_t* samplesSizes, unsigned nbSamples,
@@ -614,7 +614,7 @@ ZDICT_trainFromBuffer_fastCover(void* dictBuffer, size_t dictBufferCapacity,
}
ZDICTLIB_API size_t
ZDICTLIB_STATIC_API size_t
ZDICT_optimizeTrainFromBuffer_fastCover(
void* dictBuffer, size_t dictBufferCapacity,
const void* samplesBuffer,
Utilities/cmzstd/lib/dictBuilder/zdict.c
+14 -8
View File
@@ -74,9 +74,9 @@ static const U32 g_selectivity_default = 9;
* Console display
***************************************/
#undef DISPLAY
#define DISPLAY(...) { fprintf(stderr, __VA_ARGS__); fflush( stderr ); }
#define DISPLAY(...) do { fprintf(stderr, __VA_ARGS__); fflush( stderr ); } while (0)
#undef DISPLAYLEVEL
#define DISPLAYLEVEL(l, ...) if (notificationLevel>=l) { DISPLAY(__VA_ARGS__); } /* 0 : no display; 1: errors; 2: default; 3: details; 4: debug */
#define DISPLAYLEVEL(l, ...) do { if (notificationLevel>=l) { DISPLAY(__VA_ARGS__); } } while (0) /* 0 : no display; 1: errors; 2: default; 3: details; 4: debug */
static clock_t ZDICT_clockSpan(clock_t nPrevious) { return clock() - nPrevious; }
@@ -477,10 +477,16 @@ static size_t ZDICT_trainBuffer_legacy(dictItem* dictList, U32 dictListSize,
clock_t const refreshRate = CLOCKS_PER_SEC * 3 / 10;
# undef DISPLAYUPDATE
# define DISPLAYUPDATE(l, ...) if (notificationLevel>=l) { \
if (ZDICT_clockSpan(displayClock) > refreshRate) \
{ displayClock = clock(); DISPLAY(__VA_ARGS__); \
if (notificationLevel>=4) fflush(stderr); } }
# define DISPLAYUPDATE(l, ...) \
do { \
if (notificationLevel>=l) { \
if (ZDICT_clockSpan(displayClock) > refreshRate) { \
displayClock = clock(); \
DISPLAY(__VA_ARGS__); \
} \
if (notificationLevel>=4) fflush(stderr); \
} \
} while (0)
/* init */
DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */
@@ -574,7 +580,7 @@ static void ZDICT_countEStats(EStats_ress_t esr, const ZSTD_parameters* params,
if (ZSTD_isError(cSize)) { DISPLAYLEVEL(3, "warning : could not compress sample size %u \n", (unsigned)srcSize); return; }
if (cSize) { /* if == 0; block is not compressible */
const seqStore_t* const seqStorePtr = ZSTD_getSeqStore(esr.zc);
const SeqStore_t* const seqStorePtr = ZSTD_getSeqStore(esr.zc);
/* literals stats */
{ const BYTE* bytePtr;
@@ -602,7 +608,7 @@ static void ZDICT_countEStats(EStats_ress_t esr, const ZSTD_parameters* params,
}
if (nbSeq >= 2) { /* rep offsets */
const seqDef* const seq = seqStorePtr->sequencesStart;
const SeqDef* const seq = seqStorePtr->sequencesStart;
U32 offset1 = seq[0].offBase - ZSTD_REP_NUM;
U32 offset2 = seq[1].offBase - ZSTD_REP_NUM;
if (offset1 >= MAXREPOFFSET) offset1 = 0;
Utilities/cmzstd/lib/zdict.h
+15 -8
View File
@@ -8,16 +8,16 @@
* You may select, at your option, one of the above-listed licenses.
*/
#if defined (__cplusplus)
extern "C" {
#endif
#ifndef ZSTD_ZDICT_H
#define ZSTD_ZDICT_H
/*====== Dependencies ======*/
#include <stddef.h> /* size_t */
#if defined (__cplusplus)
extern "C" {
#endif
/* ===== ZDICTLIB_API : control library symbols visibility ===== */
#ifndef ZDICTLIB_VISIBLE
@@ -248,7 +248,7 @@ typedef struct {
* is presumed that the most profitable content is at the end of the dictionary,
* since that is the cheapest to reference.
*
* `maxDictSize` must be >= max(dictContentSize, ZSTD_DICTSIZE_MIN).
* `maxDictSize` must be >= max(dictContentSize, ZDICT_DICTSIZE_MIN).
*
* @return: size of dictionary stored into `dstDictBuffer` (<= `maxDictSize`),
* or an error code, which can be tested by ZDICT_isError().
@@ -271,11 +271,19 @@ ZDICTLIB_API size_t ZDICT_getDictHeaderSize(const void* dictBuffer, size_t dictS
ZDICTLIB_API unsigned ZDICT_isError(size_t errorCode);
ZDICTLIB_API const char* ZDICT_getErrorName(size_t errorCode);
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_ZDICT_H */
#if defined(ZDICT_STATIC_LINKING_ONLY) && !defined(ZSTD_ZDICT_H_STATIC)
#define ZSTD_ZDICT_H_STATIC
#if defined (__cplusplus)
extern "C" {
#endif
/* This can be overridden externally to hide static symbols. */
#ifndef ZDICTLIB_STATIC_API
# if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
@@ -466,9 +474,8 @@ ZDICTLIB_STATIC_API
size_t ZDICT_addEntropyTablesFromBuffer(void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,
const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples);
#endif /* ZSTD_ZDICT_H_STATIC */
#if defined (__cplusplus)
}
#endif
#endif /* ZSTD_ZDICT_H_STATIC */
Utilities/cmzstd/lib/zstd.h
+365 -187
View File
File diff suppressed because it is too large. Load diff
Utilities/cmzstd/lib/zstd_errors.h
+1 -8
View File
@@ -15,10 +15,6 @@
extern "C" {
#endif
/*===== dependency =====*/
#include <stddef.h> /* size_t */
/* ===== ZSTDERRORLIB_API : control library symbols visibility ===== */
#ifndef ZSTDERRORLIB_VISIBLE
/* Backwards compatibility with old macro name */
@@ -80,6 +76,7 @@ typedef enum {
ZSTD_error_tableLog_tooLarge = 44,
ZSTD_error_maxSymbolValue_tooLarge = 46,
ZSTD_error_maxSymbolValue_tooSmall = 48,
ZSTD_error_cannotProduce_uncompressedBlock = 49,
ZSTD_error_stabilityCondition_notRespected = 50,
ZSTD_error_stage_wrong = 60,
ZSTD_error_init_missing = 62,
@@ -100,10 +97,6 @@ typedef enum {
ZSTD_error_maxCode = 120 /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
} ZSTD_ErrorCode;
/*! ZSTD_getErrorCode() :
convert a `size_t` function result into a `ZSTD_ErrorCode` enum type,
which can be used to compare with enum list published above */
ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult);
ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code); /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */