name='decompress',
header_namespace='',
visibility=['PUBLIC'],
+ headers=subdir_glob([
+ ('decompress', '*_impl.h'),
+ ]),
srcs=glob(['decompress/zstd*.c']),
deps=[
':common',
]),
)
+cxx_library(
+ name='cpu',
+ header_namespace='',
+ visibility=['PUBLIC'],
+ exported_headers=subdir_glob([
+ ('common', 'cpu.h'),
+ ]),
+)
+
cxx_library(
name='bitstream',
header_namespace='',
deps=[
':bitstream',
':compiler',
+ ':cpu',
':entropy',
':errors',
':mem',
# endif
#endif
+/* target attribute */
+#if defined(__GNUC__)
+# define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
+#else
+# define TARGET_ATTRIBUTE(target)
+#endif
+
+/* Enable runtime BMI2 dispatch based on the CPU.
+ * Enabled for clang/gcc on x86 when BMI2 isn't enabled by default.
+ */
+#ifndef DYNAMIC_BMI2
+ #if defined(__GNUC__) && (defined(__x86_64__) || defined(_M_X86)) && !defined(__BMI2__)
+ # define DYNAMIC_BMI2 1
+ #else
+ # define DYNAMIC_BMI2 0
+ #endif
+#endif
+
/* prefetch */
#if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) /* _mm_prefetch() is not defined outside of x86/x64 */
# include <mmintrin.h> /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */
--- /dev/null
+/*
+ * Copyright (c) 2018-present, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMMON_CPU_H
+#define ZSTD_COMMON_CPU_H
+
+/**
+ * Implementation taken from folly/CpuId.h
+ * https://github.com/facebook/folly/blob/master/folly/CpuId.h
+ */
+
+#include <string.h>
+
+#include "mem.h"
+
+#ifdef _MSC_VER
+#include <intrin.h>
+#endif
+
+typedef struct {
+ U32 f1c;
+ U32 f1d;
+ U32 f7b;
+ U32 f7c;
+} ZSTD_cpuid_t;
+
+MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
+ U32 f1c = 0;
+ U32 f1d = 0;
+ U32 f7b = 0;
+ U32 f7c = 0;
+#ifdef _MSC_VER
+ int reg[4];
+ __cpuid((int*)reg, 0);
+ {
+ int const n = reg[0];
+ if (n >= 1) {
+ __cpuid((int*)reg, 1);
+ f1c = (U32)reg[2];
+ f1d = (U32)reg[3];
+ }
+ if (n >= 7) {
+ __cpuidex((int*)reg, 7, 0);
+ f7b = (U32)reg[1];
+ f7c = (U32)reg[2];
+ }
+ }
+#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
+ * handle the save and restore to avoid clobbering the register
+ */
+ U32 n;
+ __asm__(
+ "pushl %%ebx\n\t"
+ "cpuid\n\t"
+ "popl %%ebx\n\t"
+ : "=a"(n)
+ : "a"(0)
+ : "ecx", "edx");
+ if (n >= 1) {
+ U32 f1a;
+ __asm__(
+ "pushl %%ebx\n\t"
+ "cpuid\n\t"
+ "popl %%ebx\n\t"
+ : "=a"(f1a), "=c"(f1c), "=d"(f1d)
+ : "a"(1)
+ :);
+ }
+ if (n >= 7) {
+ __asm__(
+ "pushl %%ebx\n\t"
+ "cpuid\n\t"
+ "movl %%ebx, %%eax\n\r"
+ "popl %%ebx"
+ : "=a"(f7b), "=c"(f7c)
+ : "a"(7), "c"(0)
+ : "edx");
+ }
+#elif defined(__x86_64__) || defined(_M_X64) || defined(__i386__)
+ U32 n;
+ __asm__("cpuid" : "=a"(n) : "a"(0) : "ebx", "ecx", "edx");
+ if (n >= 1) {
+ U32 f1a;
+ __asm__("cpuid" : "=a"(f1a), "=c"(f1c), "=d"(f1d) : "a"(1) : "ebx");
+ }
+ if (n >= 7) {
+ U32 f7a;
+ __asm__("cpuid"
+ : "=a"(f7a), "=b"(f7b), "=c"(f7c)
+ : "a"(7), "c"(0)
+ : "edx");
+ }
+#endif
+ {
+ ZSTD_cpuid_t cpuid;
+ cpuid.f1c = f1c;
+ cpuid.f1d = f1d;
+ cpuid.f7b = f7b;
+ cpuid.f7c = f7c;
+ return cpuid;
+ }
+}
+
+#define X(name, r, bit) \
+ MEM_STATIC int ZSTD_cpuid_##name(ZSTD_cpuid_t const cpuid) { \
+ return ((cpuid.r) & (1U << bit)) != 0; \
+ }
+
+/* cpuid(1): Processor Info and Feature Bits. */
+#define C(name, bit) X(name, f1c, bit)
+ C(sse3, 0)
+ C(pclmuldq, 1)
+ C(dtes64, 2)
+ C(monitor, 3)
+ C(dscpl, 4)
+ C(vmx, 5)
+ C(smx, 6)
+ C(eist, 7)
+ C(tm2, 8)
+ C(ssse3, 9)
+ C(cnxtid, 10)
+ C(fma, 12)
+ C(cx16, 13)
+ C(xtpr, 14)
+ C(pdcm, 15)
+ C(pcid, 17)
+ C(dca, 18)
+ C(sse41, 19)
+ C(sse42, 20)
+ C(x2apic, 21)
+ C(movbe, 22)
+ C(popcnt, 23)
+ C(tscdeadline, 24)
+ C(aes, 25)
+ C(xsave, 26)
+ C(osxsave, 27)
+ C(avx, 28)
+ C(f16c, 29)
+ C(rdrand, 30)
+#undef C
+#define D(name, bit) X(name, f1d, bit)
+ D(fpu, 0)
+ D(vme, 1)
+ D(de, 2)
+ D(pse, 3)
+ D(tsc, 4)
+ D(msr, 5)
+ D(pae, 6)
+ D(mce, 7)
+ D(cx8, 8)
+ D(apic, 9)
+ D(sep, 11)
+ D(mtrr, 12)
+ D(pge, 13)
+ D(mca, 14)
+ D(cmov, 15)
+ D(pat, 16)
+ D(pse36, 17)
+ D(psn, 18)
+ D(clfsh, 19)
+ D(ds, 21)
+ D(acpi, 22)
+ D(mmx, 23)
+ D(fxsr, 24)
+ D(sse, 25)
+ D(sse2, 26)
+ D(ss, 27)
+ D(htt, 28)
+ D(tm, 29)
+ D(pbe, 31)
+#undef D
+
+/* cpuid(7): Extended Features. */
+#define B(name, bit) X(name, f7b, bit)
+ B(bmi1, 3)
+ B(hle, 4)
+ B(avx2, 5)
+ B(smep, 7)
+ B(bmi2, 8)
+ B(erms, 9)
+ B(invpcid, 10)
+ B(rtm, 11)
+ B(mpx, 14)
+ B(avx512f, 16)
+ B(avx512dq, 17)
+ B(rdseed, 18)
+ B(adx, 19)
+ B(smap, 20)
+ B(avx512ifma, 21)
+ B(pcommit, 22)
+ B(clflushopt, 23)
+ B(clwb, 24)
+ B(avx512pf, 26)
+ B(avx512er, 27)
+ B(avx512cd, 28)
+ B(sha, 29)
+ B(avx512bw, 30)
+ B(avx512vl, 31)
+#undef B
+#define C(name, bit) X(name, f7c, bit)
+ C(prefetchwt1, 0)
+ C(avx512vbmi, 1)
+#undef C
+
+#undef X
+
+#endif /* ZSTD_COMMON_CPU_H */
size_t HUF_decompress1X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
size_t HUF_decompress1X4_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+/* BMI2 variants.
+ * If the CPU has BMI2 support pass bmi2=1, otherwise pass bmi2=0.
+ */
+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+size_t HUF_decompress1X2_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+
#endif /* HUF_STATIC_LINKING_ONLY */
#if defined (__cplusplus)
workSpace, sizeof(workSpace));
}
+typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX4; /* double-symbols decoding */
-static BYTE HUF_decodeSymbolX2(BIT_DStream_t* Dstream, const HUF_DEltX2* dt, const U32 dtLog)
-{
- size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */
- BYTE const c = dt[val].byte;
- BIT_skipBits(Dstream, dt[val].nbBits);
- return c;
-}
-
-#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
- *ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog)
-
-#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
- if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
- HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
-
-#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
- if (MEM_64bits()) \
- HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
-
-HINT_INLINE size_t HUF_decodeStreamX2(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX2* const dt, const U32 dtLog)
-{
- BYTE* const pStart = p;
-
- /* up to 4 symbols at a time */
- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-4)) {
- HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
- HUF_DECODE_SYMBOLX2_1(p, bitDPtr);
- HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+#define FUNCTION(fn) fn##_default
+#define TARGET
+#include "huf_decompress_impl.h"
+#undef TARGET
+#undef FUNCTION
+
+#if DYNAMIC_BMI2
+
+#define FUNCTION(fn) fn##_bmi2
+#define TARGET TARGET_ATTRIBUTE("bmi2")
+#include "huf_decompress_impl.h"
+#undef TARGET
+#undef FUNCTION
+
+#endif
+
+typedef size_t (*HUF_decompress_usingDTable_t)(void *dst, size_t dstSize,
+ const void *cSrc,
+ size_t cSrcSize,
+ const HUF_DTable *DTable);
+#if DYNAMIC_BMI2
+#define X(fn) \
+ static size_t fn(void* dst, size_t dstSize, void const* cSrc, \
+ size_t cSrcSize, HUF_DTable const* DTable, int bmi2) \
+ { \
+ if (bmi2) { \
+ return fn##_bmi2(dst, dstSize, cSrc, cSrcSize, DTable); \
+ } \
+ return fn##_default(dst, dstSize, cSrc, cSrcSize, DTable); \
}
+#else
+#define X(fn) \
+ static size_t fn(void* dst, size_t dstSize, void const* cSrc, \
+ size_t cSrcSize, HUF_DTable const* DTable, int bmi2) \
+ { \
+ (void)bmi2; \
+ return fn##_default(dst, dstSize, cSrc, cSrcSize, DTable); \
+ }
+#endif
- /* closer to the end */
- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd))
- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
-
- /* no more data to retrieve from bitstream, hence no need to reload */
- while (p < pEnd)
- HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
-
- return pEnd-pStart;
-}
-
-static size_t HUF_decompress1X2_usingDTable_internal(
- void* dst, size_t dstSize,
- const void* cSrc, size_t cSrcSize,
- const HUF_DTable* DTable)
-{
- BYTE* op = (BYTE*)dst;
- BYTE* const oend = op + dstSize;
- const void* dtPtr = DTable + 1;
- const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
- BIT_DStream_t bitD;
- DTableDesc const dtd = HUF_getDTableDesc(DTable);
- U32 const dtLog = dtd.tableLog;
-
- { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
- if (HUF_isError(errorCode)) return errorCode; }
-
- HUF_decodeStreamX2(op, &bitD, oend, dt, dtLog);
-
- /* check */
- if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
+X(HUF_decompress1X2_usingDTable_internal)
+X(HUF_decompress4X2_usingDTable_internal)
+X(HUF_decompress1X4_usingDTable_internal)
+X(HUF_decompress4X4_usingDTable_internal)
- return dstSize;
-}
+#undef X
size_t HUF_decompress1X2_usingDTable(
void* dst, size_t dstSize,
{
DTableDesc dtd = HUF_getDTableDesc(DTable);
if (dtd.tableType != 0) return ERROR(GENERIC);
- return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
+ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,
if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
ip += hSize; cSrcSize -= hSize;
- return HUF_decompress1X2_usingDTable_internal (dst, dstSize, ip, cSrcSize, DCtx);
+ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);
}
return HUF_decompress1X2_DCtx (DTable, dst, dstSize, cSrc, cSrcSize);
}
-
-static size_t HUF_decompress4X2_usingDTable_internal(
- void* dst, size_t dstSize,
- const void* cSrc, size_t cSrcSize,
- const HUF_DTable* DTable)
-{
- /* Check */
- if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
-
- { const BYTE* const istart = (const BYTE*) cSrc;
- BYTE* const ostart = (BYTE*) dst;
- BYTE* const oend = ostart + dstSize;
- const void* const dtPtr = DTable + 1;
- const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
-
- /* Init */
- BIT_DStream_t bitD1;
- BIT_DStream_t bitD2;
- BIT_DStream_t bitD3;
- BIT_DStream_t bitD4;
- size_t const length1 = MEM_readLE16(istart);
- size_t const length2 = MEM_readLE16(istart+2);
- size_t const length3 = MEM_readLE16(istart+4);
- size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
- const BYTE* const istart1 = istart + 6; /* jumpTable */
- const BYTE* const istart2 = istart1 + length1;
- const BYTE* const istart3 = istart2 + length2;
- const BYTE* const istart4 = istart3 + length3;
- const size_t segmentSize = (dstSize+3) / 4;
- BYTE* const opStart2 = ostart + segmentSize;
- BYTE* const opStart3 = opStart2 + segmentSize;
- BYTE* const opStart4 = opStart3 + segmentSize;
- BYTE* op1 = ostart;
- BYTE* op2 = opStart2;
- BYTE* op3 = opStart3;
- BYTE* op4 = opStart4;
- U32 endSignal;
- DTableDesc const dtd = HUF_getDTableDesc(DTable);
- U32 const dtLog = dtd.tableLog;
-
- if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
- { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
- if (HUF_isError(errorCode)) return errorCode; }
-
- /* 16-32 symbols per loop (4-8 symbols per stream) */
- endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
- for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; ) {
- HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
- HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
- HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
- HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
- HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
- HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
- HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
- HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
- HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
- HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
- HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
- HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
- HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
- HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
- HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
- HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
- endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
- }
-
- /* check corruption */
- if (op1 > opStart2) return ERROR(corruption_detected);
- if (op2 > opStart3) return ERROR(corruption_detected);
- if (op3 > opStart4) return ERROR(corruption_detected);
- /* note : op4 supposed already verified within main loop */
-
- /* finish bitStreams one by one */
- HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);
- HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);
- HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);
- HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog);
-
- /* check */
- endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
- if (!endSignal) return ERROR(corruption_detected);
-
- /* decoded size */
- return dstSize;
- }
-}
-
-
size_t HUF_decompress4X2_usingDTable(
void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
{
DTableDesc dtd = HUF_getDTableDesc(DTable);
if (dtd.tableType != 0) return ERROR(GENERIC);
- return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
+ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
-
-size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+static size_t HUF_decompress4X2_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
- void* workSpace, size_t wkspSize)
+ void* workSpace, size_t wkspSize, int bmi2)
{
const BYTE* ip = (const BYTE*) cSrc;
if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
ip += hSize; cSrcSize -= hSize;
- return HUF_decompress4X2_usingDTable_internal (dst, dstSize, ip, cSrcSize, dctx);
+ return HUF_decompress4X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+
+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ void* workSpace, size_t wkspSize)
+{
+ return HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, 0);
}
/* *************************/
/* double-symbols decoding */
/* *************************/
-typedef struct { U16 sequence; BYTE nbBits; BYTE length; } HUF_DEltX4; /* double-symbols decoding */
-
typedef struct { BYTE symbol; BYTE weight; } sortedSymbol_t;
/* HUF_fillDTableX4Level2() :
workSpace, sizeof(workSpace));
}
-static U32 HUF_decodeSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
-{
- size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
- memcpy(op, dt+val, 2);
- BIT_skipBits(DStream, dt[val].nbBits);
- return dt[val].length;
-}
-
-static U32 HUF_decodeLastSymbolX4(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
-{
- size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
- memcpy(op, dt+val, 1);
- if (dt[val].length==1) BIT_skipBits(DStream, dt[val].nbBits);
- else {
- if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8)) {
- BIT_skipBits(DStream, dt[val].nbBits);
- if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8))
- /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */
- DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8);
- } }
- return 1;
-}
-
-
-#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) \
- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-
-#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \
- if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-
-#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \
- if (MEM_64bits()) \
- ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-
-HINT_INLINE size_t HUF_decodeStreamX4(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd, const HUF_DEltX4* const dt, const U32 dtLog)
-{
- BYTE* const pStart = p;
-
- /* up to 8 symbols at a time */
- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-(sizeof(bitDPtr->bitContainer)-1))) {
- HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
- HUF_DECODE_SYMBOLX4_1(p, bitDPtr);
- HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
- HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
- }
-
- /* closer to end : up to 2 symbols at a time */
- while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd-2))
- HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
-
- while (p <= pEnd-2)
- HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */
-
- if (p < pEnd)
- p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog);
-
- return p-pStart;
-}
-
-
-static size_t HUF_decompress1X4_usingDTable_internal(
- void* dst, size_t dstSize,
- const void* cSrc, size_t cSrcSize,
- const HUF_DTable* DTable)
-{
- BIT_DStream_t bitD;
-
- /* Init */
- { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
- if (HUF_isError(errorCode)) return errorCode;
- }
-
- /* decode */
- { BYTE* const ostart = (BYTE*) dst;
- BYTE* const oend = ostart + dstSize;
- const void* const dtPtr = DTable+1; /* force compiler to not use strict-aliasing */
- const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;
- DTableDesc const dtd = HUF_getDTableDesc(DTable);
- HUF_decodeStreamX4(ostart, &bitD, oend, dt, dtd.tableLog);
- }
-
- /* check */
- if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
-
- /* decoded size */
- return dstSize;
-}
-
size_t HUF_decompress1X4_usingDTable(
void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
{
DTableDesc dtd = HUF_getDTableDesc(DTable);
if (dtd.tableType != 1) return ERROR(GENERIC);
- return HUF_decompress1X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
+ return HUF_decompress1X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
size_t HUF_decompress1X4_DCtx_wksp(HUF_DTable* DCtx, void* dst, size_t dstSize,
if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
ip += hSize; cSrcSize -= hSize;
- return HUF_decompress1X4_usingDTable_internal (dst, dstSize, ip, cSrcSize, DCtx);
+ return HUF_decompress1X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx, /* bmi2 */ 0);
}
return HUF_decompress1X4_DCtx(DTable, dst, dstSize, cSrc, cSrcSize);
}
-static size_t HUF_decompress4X4_usingDTable_internal(
- void* dst, size_t dstSize,
- const void* cSrc, size_t cSrcSize,
- const HUF_DTable* DTable)
-{
- if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
-
- { const BYTE* const istart = (const BYTE*) cSrc;
- BYTE* const ostart = (BYTE*) dst;
- BYTE* const oend = ostart + dstSize;
- const void* const dtPtr = DTable+1;
- const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;
-
- /* Init */
- BIT_DStream_t bitD1;
- BIT_DStream_t bitD2;
- BIT_DStream_t bitD3;
- BIT_DStream_t bitD4;
- size_t const length1 = MEM_readLE16(istart);
- size_t const length2 = MEM_readLE16(istart+2);
- size_t const length3 = MEM_readLE16(istart+4);
- size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
- const BYTE* const istart1 = istart + 6; /* jumpTable */
- const BYTE* const istart2 = istart1 + length1;
- const BYTE* const istart3 = istart2 + length2;
- const BYTE* const istart4 = istart3 + length3;
- size_t const segmentSize = (dstSize+3) / 4;
- BYTE* const opStart2 = ostart + segmentSize;
- BYTE* const opStart3 = opStart2 + segmentSize;
- BYTE* const opStart4 = opStart3 + segmentSize;
- BYTE* op1 = ostart;
- BYTE* op2 = opStart2;
- BYTE* op3 = opStart3;
- BYTE* op4 = opStart4;
- U32 endSignal;
- DTableDesc const dtd = HUF_getDTableDesc(DTable);
- U32 const dtLog = dtd.tableLog;
-
- if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
- { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
- if (HUF_isError(errorCode)) return errorCode; }
- { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
- if (HUF_isError(errorCode)) return errorCode; }
-
- /* 16-32 symbols per loop (4-8 symbols per stream) */
- endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
- for ( ; (endSignal==BIT_DStream_unfinished) & (op4<(oend-(sizeof(bitD4.bitContainer)-1))) ; ) {
- HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
- HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
- HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
- HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
- HUF_DECODE_SYMBOLX4_1(op1, &bitD1);
- HUF_DECODE_SYMBOLX4_1(op2, &bitD2);
- HUF_DECODE_SYMBOLX4_1(op3, &bitD3);
- HUF_DECODE_SYMBOLX4_1(op4, &bitD4);
- HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
- HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
- HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
- HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
- HUF_DECODE_SYMBOLX4_0(op1, &bitD1);
- HUF_DECODE_SYMBOLX4_0(op2, &bitD2);
- HUF_DECODE_SYMBOLX4_0(op3, &bitD3);
- HUF_DECODE_SYMBOLX4_0(op4, &bitD4);
-
- endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
- }
-
- /* check corruption */
- if (op1 > opStart2) return ERROR(corruption_detected);
- if (op2 > opStart3) return ERROR(corruption_detected);
- if (op3 > opStart4) return ERROR(corruption_detected);
- /* note : op4 already verified within main loop */
-
- /* finish bitStreams one by one */
- HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog);
- HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog);
- HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog);
- HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog);
-
- /* check */
- { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
- if (!endCheck) return ERROR(corruption_detected); }
-
- /* decoded size */
- return dstSize;
- }
-}
-
-
size_t HUF_decompress4X4_usingDTable(
void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
{
DTableDesc dtd = HUF_getDTableDesc(DTable);
if (dtd.tableType != 1) return ERROR(GENERIC);
- return HUF_decompress4X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
+ return HUF_decompress4X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
-
-size_t HUF_decompress4X4_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+static size_t HUF_decompress4X4_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize,
const void* cSrc, size_t cSrcSize,
- void* workSpace, size_t wkspSize)
+ void* workSpace, size_t wkspSize, int bmi2)
{
const BYTE* ip = (const BYTE*) cSrc;
if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
ip += hSize; cSrcSize -= hSize;
- return HUF_decompress4X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx);
+ return HUF_decompress4X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+
+size_t HUF_decompress4X4_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ void* workSpace, size_t wkspSize)
+{
+ return HUF_decompress4X4_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, /* bmi2 */ 0);
}
const HUF_DTable* DTable)
{
DTableDesc const dtd = HUF_getDTableDesc(DTable);
- return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) :
- HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);
+ return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :
+ HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize,
const HUF_DTable* DTable)
{
DTableDesc const dtd = HUF_getDTableDesc(DTable);
- return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable) :
- HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);
+ return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0) :
+ HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, /* bmi2 */ 0);
}
return HUF_decompress1X_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize,
workSpace, sizeof(workSpace));
}
+
+
+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)
+{
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :
+ HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+}
+
+size_t HUF_decompress1X2_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)
+{
+ const BYTE* ip = (const BYTE*) cSrc;
+
+ size_t const hSize = HUF_readDTableX2_wksp(dctx, cSrc, cSrcSize, workSpace, wkspSize);
+ if (HUF_isError(hSize)) return hSize;
+ if (hSize >= cSrcSize) return ERROR(srcSize_wrong);
+ ip += hSize; cSrcSize -= hSize;
+
+ return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx, bmi2);
+}
+
+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2)
+{
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2) :
+ HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable, bmi2);
+}
+
+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2)
+{
+ /* validation checks */
+ if (dstSize == 0) return ERROR(dstSize_tooSmall);
+ if (cSrcSize == 0) return ERROR(corruption_detected);
+
+ { U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
+ return algoNb ? HUF_decompress4X4_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2) :
+ HUF_decompress4X2_DCtx_wksp_bmi2(dctx, dst, dstSize, cSrc, cSrcSize, workSpace, wkspSize, bmi2);
+ }
+}
--- /dev/null
+/*
+ * Copyright (c) 2018-present, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef FUNCTION
+# error "FUNCTION(name) must be defined"
+#endif
+
+#ifndef TARGET
+# error "TARGET must be defined"
+#endif
+
+
+static TARGET BYTE FUNCTION(HUF_decodeSymbolX2)(BIT_DStream_t* Dstream, const HUF_DEltX2* dt, const U32 dtLog)
+{
+ size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */
+ BYTE const c = dt[val].byte;
+ BIT_skipBits(Dstream, dt[val].nbBits);
+ return c;
+}
+
+#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) \
+ *ptr++ = FUNCTION(HUF_decodeSymbolX2)(DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr) \
+ if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
+ HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
+
+#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
+ if (MEM_64bits()) \
+ HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
+
+HINT_INLINE TARGET size_t FUNCTION(HUF_decodeStreamX2)(BYTE* p, BIT_DStream_t* const bitDPtr, BYTE* const pEnd, const HUF_DEltX2* const dt, const U32 dtLog)
+{
+ BYTE* const pStart = p;
+
+ /* up to 4 symbols at a time */
+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd-4)) {
+ HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
+ HUF_DECODE_SYMBOLX2_1(p, bitDPtr);
+ HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+ }
+
+ /* closer to the end */
+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd))
+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+
+ /* no more data to retrieve from bitstream, hence no need to reload */
+ while (p < pEnd)
+ HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
+
+ return pEnd-pStart;
+}
+
+static TARGET size_t FUNCTION(HUF_decompress1X2_usingDTable_internal)(
+ void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ const HUF_DTable* DTable)
+{
+ BYTE* op = (BYTE*)dst;
+ BYTE* const oend = op + dstSize;
+ const void* dtPtr = DTable + 1;
+ const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
+ BIT_DStream_t bitD;
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ U32 const dtLog = dtd.tableLog;
+
+ { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
+ if (HUF_isError(errorCode)) return errorCode; }
+
+ FUNCTION(HUF_decodeStreamX2)(op, &bitD, oend, dt, dtLog);
+
+ /* check */
+ if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
+
+ return dstSize;
+}
+
+static TARGET size_t
+FUNCTION(HUF_decompress4X2_usingDTable_internal)(
+ void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ const HUF_DTable* DTable)
+{
+ /* Check */
+ if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
+
+ { const BYTE* const istart = (const BYTE*) cSrc;
+ BYTE* const ostart = (BYTE*) dst;
+ BYTE* const oend = ostart + dstSize;
+ const void* const dtPtr = DTable + 1;
+ const HUF_DEltX2* const dt = (const HUF_DEltX2*)dtPtr;
+
+ /* Init */
+ BIT_DStream_t bitD1;
+ BIT_DStream_t bitD2;
+ BIT_DStream_t bitD3;
+ BIT_DStream_t bitD4;
+ size_t const length1 = MEM_readLE16(istart);
+ size_t const length2 = MEM_readLE16(istart+2);
+ size_t const length3 = MEM_readLE16(istart+4);
+ size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
+ const BYTE* const istart1 = istart + 6; /* jumpTable */
+ const BYTE* const istart2 = istart1 + length1;
+ const BYTE* const istart3 = istart2 + length2;
+ const BYTE* const istart4 = istart3 + length3;
+ const size_t segmentSize = (dstSize+3) / 4;
+ BYTE* const opStart2 = ostart + segmentSize;
+ BYTE* const opStart3 = opStart2 + segmentSize;
+ BYTE* const opStart4 = opStart3 + segmentSize;
+ BYTE* op1 = ostart;
+ BYTE* op2 = opStart2;
+ BYTE* op3 = opStart3;
+ BYTE* op4 = opStart4;
+ U32 endSignal;
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ U32 const dtLog = dtd.tableLog;
+
+ if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
+ { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
+ if (HUF_isError(errorCode)) return errorCode; }
+
+ /* 16-32 symbols per loop (4-8 symbols per stream) */
+ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
+ for ( ; (endSignal==BIT_DStream_unfinished) && (op4<(oend-7)) ; ) {
+ HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+ HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+ HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+ HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+ HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
+ HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
+ HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
+ HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
+ HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
+ HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
+ HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
+ HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
+ HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
+ HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
+ HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
+ HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
+ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
+ }
+
+ /* check corruption */
+ if (op1 > opStart2) return ERROR(corruption_detected);
+ if (op2 > opStart3) return ERROR(corruption_detected);
+ if (op3 > opStart4) return ERROR(corruption_detected);
+ /* note : op4 supposed already verified within main loop */
+
+ /* finish bitStreams one by one */
+ FUNCTION(HUF_decodeStreamX2)(op1, &bitD1, opStart2, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX2)(op2, &bitD2, opStart3, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX2)(op3, &bitD3, opStart4, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX2)(op4, &bitD4, oend, dt, dtLog);
+
+ /* check */
+ endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
+ if (!endSignal) return ERROR(corruption_detected);
+
+ /* decoded size */
+ return dstSize;
+ }
+}
+
+#undef HUF_DECODE_SYMBOLX2_0
+#undef HUF_DECODE_SYMBOLX2_1
+#undef HUF_DECODE_SYMBOLX2_2
+
+static TARGET U32 FUNCTION(HUF_decodeSymbolX4)(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
+{
+ size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
+ memcpy(op, dt+val, 2);
+ BIT_skipBits(DStream, dt[val].nbBits);
+ return dt[val].length;
+}
+
+static TARGET U32 FUNCTION(HUF_decodeLastSymbolX4)(void* op, BIT_DStream_t* DStream, const HUF_DEltX4* dt, const U32 dtLog)
+{
+ size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
+ memcpy(op, dt+val, 1);
+ if (dt[val].length==1) BIT_skipBits(DStream, dt[val].nbBits);
+ else {
+ if (DStream->bitsConsumed < (sizeof(DStream->bitContainer)*8)) {
+ BIT_skipBits(DStream, dt[val].nbBits);
+ if (DStream->bitsConsumed > (sizeof(DStream->bitContainer)*8))
+ /* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */
+ DStream->bitsConsumed = (sizeof(DStream->bitContainer)*8);
+ } }
+ return 1;
+}
+
+
+#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) \
+ ptr += FUNCTION(HUF_decodeSymbolX4)(ptr, DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr) \
+ if (MEM_64bits() || (HUF_TABLELOG_MAX<=12)) \
+ ptr += FUNCTION(HUF_decodeSymbolX4)(ptr, DStreamPtr, dt, dtLog)
+
+#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \
+ if (MEM_64bits()) \
+ ptr += FUNCTION(HUF_decodeSymbolX4)(ptr, DStreamPtr, dt, dtLog)
+
+HINT_INLINE TARGET size_t FUNCTION(HUF_decodeStreamX4)(BYTE* p, BIT_DStream_t* bitDPtr, BYTE* const pEnd, const HUF_DEltX4* const dt, const U32 dtLog)
+{
+ BYTE* const pStart = p;
+
+ /* up to 8 symbols at a time */
+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd-(sizeof(bitDPtr->bitContainer)-1))) {
+ HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
+ HUF_DECODE_SYMBOLX4_1(p, bitDPtr);
+ HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
+ HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
+ }
+
+ /* closer to end : up to 2 symbols at a time */
+ while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd-2))
+ HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
+
+ while (p <= pEnd-2)
+ HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */
+
+ if (p < pEnd)
+ p += FUNCTION(HUF_decodeLastSymbolX4)(p, bitDPtr, dt, dtLog);
+
+ return p-pStart;
+}
+
+static TARGET size_t FUNCTION(HUF_decompress1X4_usingDTable_internal)(
+ void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ const HUF_DTable* DTable)
+{
+ BIT_DStream_t bitD;
+
+ /* Init */
+ { size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
+ if (HUF_isError(errorCode)) return errorCode;
+ }
+
+ /* decode */
+ { BYTE* const ostart = (BYTE*) dst;
+ BYTE* const oend = ostart + dstSize;
+ const void* const dtPtr = DTable+1; /* force compiler to not use strict-aliasing */
+ const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ FUNCTION(HUF_decodeStreamX4)(ostart, &bitD, oend, dt, dtd.tableLog);
+ }
+
+ /* check */
+ if (!BIT_endOfDStream(&bitD)) return ERROR(corruption_detected);
+
+ /* decoded size */
+ return dstSize;
+}
+
+static TARGET size_t FUNCTION(HUF_decompress4X4_usingDTable_internal)(
+ void* dst, size_t dstSize,
+ const void* cSrc, size_t cSrcSize,
+ const HUF_DTable* DTable)
+{
+ if (cSrcSize < 10) return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
+
+ { const BYTE* const istart = (const BYTE*) cSrc;
+ BYTE* const ostart = (BYTE*) dst;
+ BYTE* const oend = ostart + dstSize;
+ const void* const dtPtr = DTable+1;
+ const HUF_DEltX4* const dt = (const HUF_DEltX4*)dtPtr;
+
+ /* Init */
+ BIT_DStream_t bitD1;
+ BIT_DStream_t bitD2;
+ BIT_DStream_t bitD3;
+ BIT_DStream_t bitD4;
+ size_t const length1 = MEM_readLE16(istart);
+ size_t const length2 = MEM_readLE16(istart+2);
+ size_t const length3 = MEM_readLE16(istart+4);
+ size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
+ const BYTE* const istart1 = istart + 6; /* jumpTable */
+ const BYTE* const istart2 = istart1 + length1;
+ const BYTE* const istart3 = istart2 + length2;
+ const BYTE* const istart4 = istart3 + length3;
+ size_t const segmentSize = (dstSize+3) / 4;
+ BYTE* const opStart2 = ostart + segmentSize;
+ BYTE* const opStart3 = opStart2 + segmentSize;
+ BYTE* const opStart4 = opStart3 + segmentSize;
+ BYTE* op1 = ostart;
+ BYTE* op2 = opStart2;
+ BYTE* op3 = opStart3;
+ BYTE* op4 = opStart4;
+ U32 endSignal;
+ DTableDesc const dtd = HUF_getDTableDesc(DTable);
+ U32 const dtLog = dtd.tableLog;
+
+ if (length4 > cSrcSize) return ERROR(corruption_detected); /* overflow */
+ { size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
+ if (HUF_isError(errorCode)) return errorCode; }
+ { size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
+ if (HUF_isError(errorCode)) return errorCode; }
+
+ /* 16-32 symbols per loop (4-8 symbols per stream) */
+ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
+ for ( ; (endSignal==BIT_DStream_unfinished) & (op4<(oend-(sizeof(bitD4.bitContainer)-1))) ; ) {
+ HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
+ HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
+ HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
+ HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
+ HUF_DECODE_SYMBOLX4_1(op1, &bitD1);
+ HUF_DECODE_SYMBOLX4_1(op2, &bitD2);
+ HUF_DECODE_SYMBOLX4_1(op3, &bitD3);
+ HUF_DECODE_SYMBOLX4_1(op4, &bitD4);
+ HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
+ HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
+ HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
+ HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
+ HUF_DECODE_SYMBOLX4_0(op1, &bitD1);
+ HUF_DECODE_SYMBOLX4_0(op2, &bitD2);
+ HUF_DECODE_SYMBOLX4_0(op3, &bitD3);
+ HUF_DECODE_SYMBOLX4_0(op4, &bitD4);
+
+ endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
+ }
+
+ /* check corruption */
+ if (op1 > opStart2) return ERROR(corruption_detected);
+ if (op2 > opStart3) return ERROR(corruption_detected);
+ if (op3 > opStart4) return ERROR(corruption_detected);
+ /* note : op4 already verified within main loop */
+
+ /* finish bitStreams one by one */
+ FUNCTION(HUF_decodeStreamX4)(op1, &bitD1, opStart2, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX4)(op2, &bitD2, opStart3, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX4)(op3, &bitD3, opStart4, dt, dtLog);
+ FUNCTION(HUF_decodeStreamX4)(op4, &bitD4, oend, dt, dtLog);
+
+ /* check */
+ { U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
+ if (!endCheck) return ERROR(corruption_detected); }
+
+ /* decoded size */
+ return dstSize;
+ }
+}
+
+#undef HUF_DECODE_SYMBOLX4_0
+#undef HUF_DECODE_SYMBOLX4_1
+#undef HUF_DECODE_SYMBOLX4_2
* Dependencies
*********************************************************/
#include <string.h> /* memcpy, memmove, memset */
+#include "cpu.h"
#include "mem.h" /* low level memory routines */
#define FSE_STATIC_LINKING_ONLY
#include "fse.h"
size_t litSize;
size_t rleSize;
size_t staticSize;
+ int bmi2;
/* streaming */
ZSTD_DDict* ddictLocal;
dctx->inBuffSize = 0;
dctx->outBuffSize = 0;
dctx->streamStage = zdss_init;
+ dctx->bmi2 = ZSTD_cpuid_bmi2(ZSTD_cpuid());
}
ZSTD_DCtx* ZSTD_initStaticDCtx(void *workspace, size_t workspaceSize)
if (HUF_isError((litEncType==set_repeat) ?
( singleStream ?
- HUF_decompress1X_usingDTable(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr) :
- HUF_decompress4X_usingDTable(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr) ) :
+ HUF_decompress1X_usingDTable_bmi2(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr, dctx->bmi2) :
+ HUF_decompress4X_usingDTable_bmi2(dctx->litBuffer, litSize, istart+lhSize, litCSize, dctx->HUFptr, dctx->bmi2) ) :
( singleStream ?
- HUF_decompress1X2_DCtx_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize,
- dctx->entropy.workspace, sizeof(dctx->entropy.workspace)) :
- HUF_decompress4X_hufOnly_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize,
- dctx->entropy.workspace, sizeof(dctx->entropy.workspace)))))
+ HUF_decompress1X2_DCtx_wksp_bmi2(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize,
+ dctx->entropy.workspace, sizeof(dctx->entropy.workspace), dctx->bmi2) :
+ HUF_decompress4X_hufOnly_wksp_bmi2(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart+lhSize, litCSize,
+ dctx->entropy.workspace, sizeof(dctx->entropy.workspace), dctx->bmi2))))
return ERROR(corruption_detected);
dctx->litPtr = dctx->litBuffer;
}
-static void
-ZSTD_updateFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD)
-{
- ZSTD_seqSymbol const DInfo = DStatePtr->table[DStatePtr->state];
- U32 const nbBits = DInfo.nbBits;
- size_t const lowBits = BIT_readBits(bitD, nbBits);
- DStatePtr->state = DInfo.nextState + lowBits;
-}
-
-
-typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
-
-/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum
- * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)
- * bits before reloading. This value is the maximum number of bytes we read
- * after reloading when we are decoding long offets.
- */
-#define LONG_OFFSETS_MAX_EXTRA_BITS_32 \
- (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32 \
- ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32 \
- : 0)
-
-static seq_t ZSTD_decodeSequence(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
-{
- seq_t seq;
- U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;
- U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;
- U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;
- U32 const totalBits = llBits+mlBits+ofBits;
- U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;
- U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;
- U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;
-
- /* sequence */
- { size_t offset;
- if (!ofBits)
- offset = 0;
- else {
- ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
- ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
- assert(ofBits <= MaxOff);
- if (MEM_32bits() && longOffsets && (ofBits >= STREAM_ACCUMULATOR_MIN_32)) {
- U32 const extraBits = ofBits - MIN(ofBits, 32 - seqState->DStream.bitsConsumed);
- offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
- BIT_reloadDStream(&seqState->DStream);
- if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
- assert(extraBits <= LONG_OFFSETS_MAX_EXTRA_BITS_32); /* to avoid another reload */
- } else {
- offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits/*>0*/); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */
- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
- }
- }
-
- if (ofBits <= 1) {
- offset += (llBase==0);
- if (offset) {
- 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 */
- if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
- seqState->prevOffset[1] = seqState->prevOffset[0];
- seqState->prevOffset[0] = offset = temp;
- } else { /* offset == 0 */
- offset = seqState->prevOffset[0];
- }
- } else {
- seqState->prevOffset[2] = seqState->prevOffset[1];
- seqState->prevOffset[1] = seqState->prevOffset[0];
- seqState->prevOffset[0] = offset;
- }
- seq.offset = offset;
- }
-
- seq.matchLength = mlBase
- + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits/*>0*/) : 0); /* <= 16 bits */
- if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
- BIT_reloadDStream(&seqState->DStream);
- if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
- BIT_reloadDStream(&seqState->DStream);
- /* Ensure there are enough bits to read the rest of data in 64-bit mode. */
- ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
-
- seq.litLength = llBase
- + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits/*>0*/) : 0); /* <= 16 bits */
- if (MEM_32bits())
- BIT_reloadDStream(&seqState->DStream);
-
- DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
- (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
-
- /* ANS state update */
- ZSTD_updateFseState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */
- ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */
- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
- ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */
-
- return seq;
-}
-
-
HINT_INLINE
size_t ZSTD_execSequence(BYTE* op,
BYTE* const oend, seq_t sequence,
}
-static void
-ZSTD_initFseState(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt)
-{
- const void* ptr = dt;
- const ZSTD_seqSymbol_header* const DTableH = (const ZSTD_seqSymbol_header*)ptr;
- DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
- DEBUGLOG(6, "ZSTD_initFseState : val=%u using %u bits",
- (U32)DStatePtr->state, DTableH->tableLog);
- BIT_reloadDStream(bitD);
- DStatePtr->table = dt + 1;
-}
-
-static size_t ZSTD_decompressSequences(
- ZSTD_DCtx* dctx,
- void* dst, size_t maxDstSize,
- const void* seqStart, size_t seqSize,
- const ZSTD_longOffset_e isLongOffset)
-{
- const BYTE* ip = (const BYTE*)seqStart;
- const BYTE* const iend = ip + seqSize;
- BYTE* const ostart = (BYTE* const)dst;
- BYTE* const oend = ostart + maxDstSize;
- BYTE* op = ostart;
- const BYTE* litPtr = dctx->litPtr;
- const BYTE* const litEnd = litPtr + dctx->litSize;
- const BYTE* const base = (const BYTE*) (dctx->base);
- const BYTE* const vBase = (const BYTE*) (dctx->vBase);
- const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
- int nbSeq;
- DEBUGLOG(5, "ZSTD_decompressSequences");
-
- /* Build Decoding Tables */
- { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
- DEBUGLOG(5, "ZSTD_decodeSeqHeaders: size=%u, nbSeq=%i",
- (U32)seqHSize, nbSeq);
- if (ZSTD_isError(seqHSize)) return seqHSize;
- ip += seqHSize;
- }
-
- /* Regen sequences */
- if (nbSeq) {
- seqState_t seqState;
- dctx->fseEntropy = 1;
- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
- CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
- ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
- ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
- ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-
- for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {
- nbSeq--;
- { seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
- size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
- DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
- op += oneSeqSize;
- } }
-
- /* check if reached exact end */
- DEBUGLOG(5, "ZSTD_decompressSequences: after decode loop, remaining nbSeq : %i", nbSeq);
- if (nbSeq) return ERROR(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;
- if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
- memcpy(op, litPtr, lastLLSize);
- op += lastLLSize;
- }
-
- return op-ostart;
-}
-
-
-
-HINT_INLINE
-seq_t ZSTD_decodeSequenceLong(seqState_t* seqState, ZSTD_longOffset_e const longOffsets)
-{
- seq_t seq;
- U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;
- U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;
- U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;
- U32 const totalBits = llBits+mlBits+ofBits;
- U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;
- U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;
- U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;
-
- /* sequence */
- { size_t offset;
- if (!ofBits)
- offset = 0;
- else {
- ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
- ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
- assert(ofBits <= MaxOff);
- if (MEM_32bits() && longOffsets) {
- U32 const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN_32-1);
- offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
- if (MEM_32bits() || extraBits) BIT_reloadDStream(&seqState->DStream);
- if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
- } else {
- offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */
- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
- }
- }
-
- if (ofBits <= 1) {
- offset += (llBase==0);
- if (offset) {
- 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 */
- if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
- seqState->prevOffset[1] = seqState->prevOffset[0];
- seqState->prevOffset[0] = offset = temp;
- } else {
- offset = seqState->prevOffset[0];
- }
- } else {
- seqState->prevOffset[2] = seqState->prevOffset[1];
- seqState->prevOffset[1] = seqState->prevOffset[0];
- seqState->prevOffset[0] = offset;
- }
- seq.offset = offset;
- }
-
- seq.matchLength = mlBase + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */
- if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
- BIT_reloadDStream(&seqState->DStream);
- if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
- BIT_reloadDStream(&seqState->DStream);
- /* Verify that there is enough bits to read the rest of the data in 64-bit mode. */
- ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
-
- seq.litLength = llBase + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */
- if (MEM_32bits())
- BIT_reloadDStream(&seqState->DStream);
-
- { size_t const pos = seqState->pos + seq.litLength;
- const BYTE* const matchBase = (seq.offset > pos) ? seqState->dictEnd : seqState->prefixStart;
- seq.match = matchBase + pos - seq.offset; /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
- * No consequence though : no memory access will occur, overly large offset will be detected in ZSTD_execSequenceLong() */
- seqState->pos = pos + seq.matchLength;
- }
-
- /* ANS state update */
- ZSTD_updateFseState(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */
- ZSTD_updateFseState(&seqState->stateML, &seqState->DStream); /* <= 9 bits */
- if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
- ZSTD_updateFseState(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */
-
- return seq;
-}
-
-
HINT_INLINE
size_t ZSTD_execSequenceLong(BYTE* op,
BYTE* const oend, seq_t sequence,
return sequenceLength;
}
-static size_t ZSTD_decompressSequencesLong(
- ZSTD_DCtx* dctx,
- void* dst, size_t maxDstSize,
- const void* seqStart, size_t seqSize,
- const ZSTD_longOffset_e isLongOffset)
-{
- const BYTE* ip = (const BYTE*)seqStart;
- const BYTE* const iend = ip + seqSize;
- BYTE* const ostart = (BYTE* const)dst;
- BYTE* const oend = ostart + maxDstSize;
- BYTE* op = ostart;
- const BYTE* litPtr = dctx->litPtr;
- const BYTE* const litEnd = litPtr + dctx->litSize;
- const BYTE* const prefixStart = (const BYTE*) (dctx->base);
- const BYTE* const dictStart = (const BYTE*) (dctx->vBase);
- const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
- int nbSeq;
-
- /* Build Decoding Tables */
- { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
- if (ZSTD_isError(seqHSize)) return seqHSize;
- ip += seqHSize;
- }
+typedef enum { ZSTD_lo_isRegularOffset, ZSTD_lo_isLongOffset=1 } ZSTD_longOffset_e;
- /* Regen sequences */
- if (nbSeq) {
-#define STORED_SEQS 4
-#define STOSEQ_MASK (STORED_SEQS-1)
-#define ADVANCED_SEQS 4
- seq_t sequences[STORED_SEQS];
- int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);
- seqState_t seqState;
- int seqNb;
- dctx->fseEntropy = 1;
- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
- seqState.prefixStart = prefixStart;
- seqState.pos = (size_t)(op-prefixStart);
- seqState.dictEnd = dictEnd;
- CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
- ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
- ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
- ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-
- /* prepare in advance */
- for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
- sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, isLongOffset);
- }
- if (seqNb<seqAdvance) return ERROR(corruption_detected);
-
- /* decode and decompress */
- for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {
- seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, isLongOffset);
- size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
- PREFETCH(sequence.match); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
- sequences[seqNb&STOSEQ_MASK] = sequence;
- op += oneSeqSize;
- }
- if (seqNb<nbSeq) return ERROR(corruption_detected);
-
- /* finish queue */
- seqNb -= seqAdvance;
- for ( ; seqNb<nbSeq ; seqNb++) {
- size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STOSEQ_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
- if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
- op += oneSeqSize;
- }
+#define FUNCTION(fn) fn##_default
+#define TARGET
+#include "zstd_decompress_impl.h"
+#undef TARGET
+#undef FUNCTION
- /* save reps for next block */
- { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
- }
+#if DYNAMIC_BMI2
- /* last literal segment */
- { size_t const lastLLSize = litEnd - litPtr;
- if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
- memcpy(op, litPtr, lastLLSize);
- op += lastLLSize;
- }
+#define FUNCTION(fn) fn##_bmi2
+#define TARGET TARGET_ATTRIBUTE("bmi2")
+#include "zstd_decompress_impl.h"
+#undef TARGET
+#undef FUNCTION
- return op-ostart;
+#endif
+
+typedef size_t (*ZSTD_decompressSequences_t)(
+ ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart,
+ size_t seqSize, const ZSTD_longOffset_e isLongOffset);
+
+static size_t ZSTD_decompressSequences(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
+ const void* seqStart, size_t seqSize,
+ const ZSTD_longOffset_e isLongOffset)
+{
+#if DYNAMIC_BMI2
+ if (dctx->bmi2) {
+ return ZSTD_decompressSequences_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, isLongOffset);
+ }
+#endif
+ return ZSTD_decompressSequences_default(dctx, dst, maxDstSize, seqStart, seqSize, isLongOffset);
}
+static size_t ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx, void* dst, size_t maxDstSize,
+ const void* seqStart, size_t seqSize,
+ const ZSTD_longOffset_e isLongOffset)
+{
+#if DYNAMIC_BMI2
+ if (dctx->bmi2) {
+ return ZSTD_decompressSequencesLong_bmi2(dctx, dst, maxDstSize, seqStart, seqSize, isLongOffset);
+ }
+#endif
+ return ZSTD_decompressSequencesLong_default(dctx, dst, maxDstSize, seqStart, seqSize, isLongOffset);
+}
static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
void* dst, size_t dstCapacity,
--- /dev/null
+/*
+ * Copyright (c) 2018-present, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef FUNCTION
+# error "FUNCTION(name) must be defined"
+#endif
+
+#ifndef TARGET
+# error "TARGET must be defined"
+#endif
+
+static TARGET void
+FUNCTION(ZSTD_updateFseState)(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD)
+{
+ ZSTD_seqSymbol const DInfo = DStatePtr->table[DStatePtr->state];
+ U32 const nbBits = DInfo.nbBits;
+ size_t const lowBits = BIT_readBits(bitD, nbBits);
+ DStatePtr->state = DInfo.nextState + lowBits;
+}
+
+/* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum
+ * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)
+ * bits before reloading. This value is the maximum number of bytes we read
+ * after reloading when we are decoding long offets.
+ */
+#define LONG_OFFSETS_MAX_EXTRA_BITS_32 \
+ (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32 \
+ ? ZSTD_WINDOWLOG_MAX_32 - STREAM_ACCUMULATOR_MIN_32 \
+ : 0)
+
+static TARGET seq_t
+FUNCTION(ZSTD_decodeSequence)(seqState_t* seqState, const ZSTD_longOffset_e longOffsets)
+{
+ seq_t seq;
+ U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;
+ U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;
+ U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;
+ U32 const totalBits = llBits+mlBits+ofBits;
+ U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;
+ U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;
+ U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;
+
+ /* sequence */
+ { size_t offset;
+ if (!ofBits)
+ offset = 0;
+ else {
+ ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
+ ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
+ assert(ofBits <= MaxOff);
+ if (MEM_32bits() && longOffsets && (ofBits >= STREAM_ACCUMULATOR_MIN_32)) {
+ U32 const extraBits = ofBits - MIN(ofBits, 32 - seqState->DStream.bitsConsumed);
+ offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
+ BIT_reloadDStream(&seqState->DStream);
+ if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
+ assert(extraBits <= LONG_OFFSETS_MAX_EXTRA_BITS_32); /* to avoid another reload */
+ } else {
+ offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits/*>0*/); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */
+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
+ }
+ }
+
+ if (ofBits <= 1) {
+ offset += (llBase==0);
+ if (offset) {
+ 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 */
+ if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
+ seqState->prevOffset[1] = seqState->prevOffset[0];
+ seqState->prevOffset[0] = offset = temp;
+ } else { /* offset == 0 */
+ offset = seqState->prevOffset[0];
+ }
+ } else {
+ seqState->prevOffset[2] = seqState->prevOffset[1];
+ seqState->prevOffset[1] = seqState->prevOffset[0];
+ seqState->prevOffset[0] = offset;
+ }
+ seq.offset = offset;
+ }
+
+ seq.matchLength = mlBase
+ + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits/*>0*/) : 0); /* <= 16 bits */
+ if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
+ BIT_reloadDStream(&seqState->DStream);
+ if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
+ BIT_reloadDStream(&seqState->DStream);
+ /* Ensure there are enough bits to read the rest of data in 64-bit mode. */
+ ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
+
+ seq.litLength = llBase
+ + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits/*>0*/) : 0); /* <= 16 bits */
+ if (MEM_32bits())
+ BIT_reloadDStream(&seqState->DStream);
+
+ DEBUGLOG(6, "seq: litL=%u, matchL=%u, offset=%u",
+ (U32)seq.litLength, (U32)seq.matchLength, (U32)seq.offset);
+
+ /* ANS state update */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateML, &seqState->DStream); /* <= 9 bits */
+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */
+
+ return seq;
+}
+
+
+
+HINT_INLINE seq_t
+FUNCTION(ZSTD_decodeSequenceLong)(seqState_t* seqState, ZSTD_longOffset_e const longOffsets)
+{
+ seq_t seq;
+ U32 const llBits = seqState->stateLL.table[seqState->stateLL.state].nbAdditionalBits;
+ U32 const mlBits = seqState->stateML.table[seqState->stateML.state].nbAdditionalBits;
+ U32 const ofBits = seqState->stateOffb.table[seqState->stateOffb.state].nbAdditionalBits;
+ U32 const totalBits = llBits+mlBits+ofBits;
+ U32 const llBase = seqState->stateLL.table[seqState->stateLL.state].baseValue;
+ U32 const mlBase = seqState->stateML.table[seqState->stateML.state].baseValue;
+ U32 const ofBase = seqState->stateOffb.table[seqState->stateOffb.state].baseValue;
+
+ /* sequence */
+ { size_t offset;
+ if (!ofBits)
+ offset = 0;
+ else {
+ ZSTD_STATIC_ASSERT(ZSTD_lo_isLongOffset == 1);
+ ZSTD_STATIC_ASSERT(LONG_OFFSETS_MAX_EXTRA_BITS_32 == 5);
+ assert(ofBits <= MaxOff);
+ if (MEM_32bits() && longOffsets) {
+ U32 const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN_32-1);
+ offset = ofBase + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
+ if (MEM_32bits() || extraBits) BIT_reloadDStream(&seqState->DStream);
+ if (extraBits) offset += BIT_readBitsFast(&seqState->DStream, extraBits);
+ } else {
+ offset = ofBase + BIT_readBitsFast(&seqState->DStream, ofBits); /* <= (ZSTD_WINDOWLOG_MAX-1) bits */
+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream);
+ }
+ }
+
+ if (ofBits <= 1) {
+ offset += (llBase==0);
+ if (offset) {
+ 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 */
+ if (offset != 1) seqState->prevOffset[2] = seqState->prevOffset[1];
+ seqState->prevOffset[1] = seqState->prevOffset[0];
+ seqState->prevOffset[0] = offset = temp;
+ } else {
+ offset = seqState->prevOffset[0];
+ }
+ } else {
+ seqState->prevOffset[2] = seqState->prevOffset[1];
+ seqState->prevOffset[1] = seqState->prevOffset[0];
+ seqState->prevOffset[0] = offset;
+ }
+ seq.offset = offset;
+ }
+
+ seq.matchLength = mlBase + ((mlBits>0) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <= 16 bits */
+ if (MEM_32bits() && (mlBits+llBits >= STREAM_ACCUMULATOR_MIN_32-LONG_OFFSETS_MAX_EXTRA_BITS_32))
+ BIT_reloadDStream(&seqState->DStream);
+ if (MEM_64bits() && (totalBits >= STREAM_ACCUMULATOR_MIN_64-(LLFSELog+MLFSELog+OffFSELog)))
+ BIT_reloadDStream(&seqState->DStream);
+ /* Verify that there is enough bits to read the rest of the data in 64-bit mode. */
+ ZSTD_STATIC_ASSERT(16+LLFSELog+MLFSELog+OffFSELog < STREAM_ACCUMULATOR_MIN_64);
+
+ seq.litLength = llBase + ((llBits>0) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <= 16 bits */
+ if (MEM_32bits())
+ BIT_reloadDStream(&seqState->DStream);
+
+ { size_t const pos = seqState->pos + seq.litLength;
+ const BYTE* const matchBase = (seq.offset > pos) ? seqState->dictEnd : seqState->prefixStart;
+ seq.match = matchBase + pos - seq.offset; /* note : this operation can overflow when seq.offset is really too large, which can only happen when input is corrupted.
+ * No consequence though : no memory access will occur, overly large offset will be detected in ZSTD_execSequenceLong() */
+ seqState->pos = pos + seq.matchLength;
+ }
+
+ /* ANS state update */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateLL, &seqState->DStream); /* <= 9 bits */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateML, &seqState->DStream); /* <= 9 bits */
+ if (MEM_32bits()) BIT_reloadDStream(&seqState->DStream); /* <= 18 bits */
+ FUNCTION(ZSTD_updateFseState)(&seqState->stateOffb, &seqState->DStream); /* <= 8 bits */
+
+ return seq;
+}
+
+
+static TARGET void
+FUNCTION(ZSTD_initFseState)(ZSTD_fseState* DStatePtr, BIT_DStream_t* bitD, const ZSTD_seqSymbol* dt)
+{
+ const void* ptr = dt;
+ const ZSTD_seqSymbol_header* const DTableH = (const ZSTD_seqSymbol_header*)ptr;
+ DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
+ DEBUGLOG(6, "ZSTD_initFseState : val=%u using %u bits",
+ (U32)DStatePtr->state, DTableH->tableLog);
+ BIT_reloadDStream(bitD);
+ DStatePtr->table = dt + 1;
+}
+
+static TARGET
+size_t FUNCTION(ZSTD_decompressSequences)(
+ ZSTD_DCtx* dctx,
+ void* dst, size_t maxDstSize,
+ const void* seqStart, size_t seqSize,
+ const ZSTD_longOffset_e isLongOffset)
+{
+ const BYTE* ip = (const BYTE*)seqStart;
+ const BYTE* const iend = ip + seqSize;
+ BYTE* const ostart = (BYTE* const)dst;
+ BYTE* const oend = ostart + maxDstSize;
+ BYTE* op = ostart;
+ const BYTE* litPtr = dctx->litPtr;
+ const BYTE* const litEnd = litPtr + dctx->litSize;
+ const BYTE* const base = (const BYTE*) (dctx->base);
+ const BYTE* const vBase = (const BYTE*) (dctx->vBase);
+ const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
+ int nbSeq;
+ DEBUGLOG(5, "ZSTD_decompressSequences");
+
+ /* Build Decoding Tables */
+ { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
+ DEBUGLOG(5, "ZSTD_decodeSeqHeaders: size=%u, nbSeq=%i",
+ (U32)seqHSize, nbSeq);
+ if (ZSTD_isError(seqHSize)) return seqHSize;
+ ip += seqHSize;
+ }
+
+ /* Regen sequences */
+ if (nbSeq) {
+ seqState_t seqState;
+ dctx->fseEntropy = 1;
+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
+ CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
+
+ for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {
+ nbSeq--;
+ { seq_t const sequence = FUNCTION(ZSTD_decodeSequence)(&seqState, isLongOffset);
+ size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
+ DEBUGLOG(6, "regenerated sequence size : %u", (U32)oneSeqSize);
+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
+ op += oneSeqSize;
+ } }
+
+ /* check if reached exact end */
+ DEBUGLOG(5, "ZSTD_decompressSequences: after decode loop, remaining nbSeq : %i", nbSeq);
+ if (nbSeq) return ERROR(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;
+ if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
+ memcpy(op, litPtr, lastLLSize);
+ op += lastLLSize;
+ }
+
+ return op-ostart;
+}
+
+static TARGET
+size_t FUNCTION(ZSTD_decompressSequencesLong)(
+ ZSTD_DCtx* dctx,
+ void* dst, size_t maxDstSize,
+ const void* seqStart, size_t seqSize,
+ const ZSTD_longOffset_e isLongOffset)
+{
+ const BYTE* ip = (const BYTE*)seqStart;
+ const BYTE* const iend = ip + seqSize;
+ BYTE* const ostart = (BYTE* const)dst;
+ BYTE* const oend = ostart + maxDstSize;
+ BYTE* op = ostart;
+ const BYTE* litPtr = dctx->litPtr;
+ const BYTE* const litEnd = litPtr + dctx->litSize;
+ const BYTE* const prefixStart = (const BYTE*) (dctx->base);
+ const BYTE* const dictStart = (const BYTE*) (dctx->vBase);
+ const BYTE* const dictEnd = (const BYTE*) (dctx->dictEnd);
+ int nbSeq;
+
+ /* Build Decoding Tables */
+ { size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
+ if (ZSTD_isError(seqHSize)) return seqHSize;
+ ip += seqHSize;
+ }
+
+ /* Regen sequences */
+ if (nbSeq) {
+#define STORED_SEQS 4
+#define STOSEQ_MASK (STORED_SEQS-1)
+#define ADVANCED_SEQS 4
+ seq_t sequences[STORED_SEQS];
+ int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);
+ seqState_t seqState;
+ int seqNb;
+ dctx->fseEntropy = 1;
+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
+ seqState.prefixStart = prefixStart;
+ seqState.pos = (size_t)(op-prefixStart);
+ seqState.dictEnd = dictEnd;
+ CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
+ FUNCTION(ZSTD_initFseState)(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
+
+ /* prepare in advance */
+ for (seqNb=0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && (seqNb<seqAdvance); seqNb++) {
+ sequences[seqNb] = FUNCTION(ZSTD_decodeSequenceLong)(&seqState, isLongOffset);
+ }
+ if (seqNb<seqAdvance) return ERROR(corruption_detected);
+
+ /* decode and decompress */
+ for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {
+ seq_t const sequence = FUNCTION(ZSTD_decodeSequenceLong)(&seqState, isLongOffset);
+ size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[(seqNb-ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
+ PREFETCH(sequence.match); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
+ sequences[seqNb&STOSEQ_MASK] = sequence;
+ op += oneSeqSize;
+ }
+ if (seqNb<nbSeq) return ERROR(corruption_detected);
+
+ /* finish queue */
+ seqNb -= seqAdvance;
+ for ( ; seqNb<nbSeq ; seqNb++) {
+ size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb&STOSEQ_MASK], &litPtr, litEnd, prefixStart, dictStart, dictEnd);
+ if (ZSTD_isError(oneSeqSize)) return oneSeqSize;
+ op += oneSeqSize;
+ }
+
+ /* save reps for next block */
+ { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
+#undef STORED_SEQS
+#undef STOSEQ_MASK
+#undef ADVANCED_SEQS
+ }
+
+ /* last literal segment */
+ { size_t const lastLLSize = litEnd - litPtr;
+ if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
+ memcpy(op, litPtr, lastLLSize);
+ op += lastLLSize;
+ }
+
+ return op-ostart;
+}
projects / thirdparty / zstd.git / commitdiff
