summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorYann Collet <Cyan4973@users.noreply.github.com>2020-10-31 01:44:21 -0700
committerGitHub <noreply@github.com>2020-10-31 01:44:21 -0700
commit664427aa8f6808411a2c04f003a9b364bdebe578 (patch)
treea8c7a2929f5e5a0e7be6f8eb662b1aaa2f66ddc2
parent644ab86b8cca85ac333cc640331a9798ce968cbb (diff)
parent0e3933edd435c54cc2e21e38f5d4ba7bf644a24e (diff)
downloadlz4-664427aa8f6808411a2c04f003a9b364bdebe578.tar.gz
lz4-664427aa8f6808411a2c04f003a9b364bdebe578.tar.bz2
lz4-664427aa8f6808411a2c04f003a9b364bdebe578.zip
Merge pull request #930 from remittor-pr/fix_msvc
Fix: The "inline" specifier do not use for LZ4_wildCopy8 and LZ4_wildCopy32
-rw-r--r--lib/lz4.c72
-rw-r--r--lib/lz4hc.c4
2 files changed, 38 insertions, 38 deletions
diff --git a/lib/lz4.c b/lib/lz4.c
index 0290ea2..6adfafd 100644
--- a/lib/lz4.c
+++ b/lib/lz4.c
@@ -143,7 +143,7 @@
# endif /* _MSC_VER */
#endif /* LZ4_FORCE_INLINE */
-/* LZ4_FORCE_O2_GCC_PPC64LE and LZ4_FORCE_O2_INLINE_GCC_PPC64LE
+/* LZ4_FORCE_O2 and LZ4_FORCE_INLINE
* gcc on ppc64le generates an unrolled SIMDized loop for LZ4_wildCopy8,
* together with a simple 8-byte copy loop as a fall-back path.
* However, this optimization hurts the decompression speed by >30%,
@@ -158,11 +158,11 @@
* of LZ4_wildCopy8 does not affect the compression speed.
*/
#if defined(__PPC64__) && defined(__LITTLE_ENDIAN__) && defined(__GNUC__) && !defined(__clang__)
-# define LZ4_FORCE_O2_GCC_PPC64LE __attribute__((optimize("O2")))
-# define LZ4_FORCE_O2_INLINE_GCC_PPC64LE __attribute__((optimize("O2"))) LZ4_FORCE_INLINE
+# define LZ4_FORCE_O2 __attribute__((optimize("O2")))
+# undef LZ4_FORCE_INLINE
+# define LZ4_FORCE_INLINE static __inline __attribute__((optimize("O2"),always_inline))
#else
-# define LZ4_FORCE_O2_GCC_PPC64LE
-# define LZ4_FORCE_O2_INLINE_GCC_PPC64LE static
+# define LZ4_FORCE_O2
#endif
#if (defined(__GNUC__) && (__GNUC__ >= 3)) || (defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 800)) || defined(__clang__)
@@ -299,7 +299,7 @@ typedef enum {
#define LZ4_memcpy(dst, src, size) memcpy(dst, src, size)
#endif
-static unsigned LZ4_isLittleEndian(void)
+LZ4_FORCE_INLINE unsigned LZ4_isLittleEndian(void)
{
const union { U32 u; BYTE c[4]; } one = { 1 }; /* don't use static : performance detrimental */
return one.c[0];
@@ -309,12 +309,12 @@ static unsigned LZ4_isLittleEndian(void)
#if defined(LZ4_FORCE_MEMORY_ACCESS) && (LZ4_FORCE_MEMORY_ACCESS==2)
/* lie to the compiler about data alignment; use with caution */
-static U16 LZ4_read16(const void* memPtr) { return *(const U16*) memPtr; }
-static U32 LZ4_read32(const void* memPtr) { return *(const U32*) memPtr; }
-static reg_t LZ4_read_ARCH(const void* memPtr) { return *(const reg_t*) memPtr; }
+LZ4_FORCE_INLINE U16 LZ4_read16(const void* memPtr) { return *(const U16*) memPtr; }
+LZ4_FORCE_INLINE U32 LZ4_read32(const void* memPtr) { return *(const U32*) memPtr; }
+LZ4_FORCE_INLINE reg_t LZ4_read_ARCH(const void* memPtr) { return *(const reg_t*) memPtr; }
-static void LZ4_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
-static void LZ4_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
+LZ4_FORCE_INLINE void LZ4_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
+LZ4_FORCE_INLINE void LZ4_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
#elif defined(LZ4_FORCE_MEMORY_ACCESS) && (LZ4_FORCE_MEMORY_ACCESS==1)
@@ -322,36 +322,36 @@ static void LZ4_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
/* currently only defined for gcc and icc */
typedef union { U16 u16; U32 u32; reg_t uArch; } __attribute__((packed)) unalign;
-static U16 LZ4_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
-static U32 LZ4_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
-static reg_t LZ4_read_ARCH(const void* ptr) { return ((const unalign*)ptr)->uArch; }
+LZ4_FORCE_INLINE U16 LZ4_read16(const void* ptr) { return ((const unalign*)ptr)->u16; }
+LZ4_FORCE_INLINE U32 LZ4_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
+LZ4_FORCE_INLINE reg_t LZ4_read_ARCH(const void* ptr) { return ((const unalign*)ptr)->uArch; }
-static void LZ4_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
-static void LZ4_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; }
+LZ4_FORCE_INLINE void LZ4_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
+LZ4_FORCE_INLINE void LZ4_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; }
#else /* safe and portable access using memcpy() */
-static U16 LZ4_read16(const void* memPtr)
+LZ4_FORCE_INLINE U16 LZ4_read16(const void* memPtr)
{
U16 val; LZ4_memcpy(&val, memPtr, sizeof(val)); return val;
}
-static U32 LZ4_read32(const void* memPtr)
+LZ4_FORCE_INLINE U32 LZ4_read32(const void* memPtr)
{
U32 val; LZ4_memcpy(&val, memPtr, sizeof(val)); return val;
}
-static reg_t LZ4_read_ARCH(const void* memPtr)
+LZ4_FORCE_INLINE reg_t LZ4_read_ARCH(const void* memPtr)
{
reg_t val; LZ4_memcpy(&val, memPtr, sizeof(val)); return val;
}
-static void LZ4_write16(void* memPtr, U16 value)
+LZ4_FORCE_INLINE void LZ4_write16(void* memPtr, U16 value)
{
LZ4_memcpy(memPtr, &value, sizeof(value));
}
-static void LZ4_write32(void* memPtr, U32 value)
+LZ4_FORCE_INLINE void LZ4_write32(void* memPtr, U32 value)
{
LZ4_memcpy(memPtr, &value, sizeof(value));
}
@@ -359,7 +359,7 @@ static void LZ4_write32(void* memPtr, U32 value)
#endif /* LZ4_FORCE_MEMORY_ACCESS */
-static U16 LZ4_readLE16(const void* memPtr)
+LZ4_FORCE_INLINE U16 LZ4_readLE16(const void* memPtr)
{
if (LZ4_isLittleEndian()) {
return LZ4_read16(memPtr);
@@ -369,7 +369,7 @@ static U16 LZ4_readLE16(const void* memPtr)
}
}
-static void LZ4_writeLE16(void* memPtr, U16 value)
+LZ4_FORCE_INLINE void LZ4_writeLE16(void* memPtr, U16 value)
{
if (LZ4_isLittleEndian()) {
LZ4_write16(memPtr, value);
@@ -381,7 +381,7 @@ static void LZ4_writeLE16(void* memPtr, U16 value)
}
/* customized variant of memcpy, which can overwrite up to 8 bytes beyond dstEnd */
-LZ4_FORCE_O2_INLINE_GCC_PPC64LE
+LZ4_FORCE_INLINE
void LZ4_wildCopy8(void* dstPtr, const void* srcPtr, void* dstEnd)
{
BYTE* d = (BYTE*)dstPtr;
@@ -410,7 +410,7 @@ static const int dec64table[8] = {0, 0, 0, -1, -4, 1, 2, 3};
#if LZ4_FAST_DEC_LOOP
-LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_FORCE_INLINE void
LZ4_memcpy_using_offset_base(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const size_t offset)
{
if (offset < 8) {
@@ -434,7 +434,7 @@ LZ4_memcpy_using_offset_base(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, con
/* customized variant of memcpy, which can overwrite up to 32 bytes beyond dstEnd
* this version copies two times 16 bytes (instead of one time 32 bytes)
* because it must be compatible with offsets >= 16. */
-LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_FORCE_INLINE void
LZ4_wildCopy32(void* dstPtr, const void* srcPtr, void* dstEnd)
{
BYTE* d = (BYTE*)dstPtr;
@@ -447,7 +447,7 @@ LZ4_wildCopy32(void* dstPtr, const void* srcPtr, void* dstEnd)
/* LZ4_memcpy_using_offset() presumes :
* - dstEnd >= dstPtr + MINMATCH
* - there is at least 8 bytes available to write after dstEnd */
-LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_FORCE_INLINE void
LZ4_memcpy_using_offset(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const size_t offset)
{
BYTE v[8];
@@ -2145,7 +2145,7 @@ LZ4_decompress_generic(
/*===== Instantiate the API decoding functions. =====*/
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_safe(const char* source, char* dest, int compressedSize, int maxDecompressedSize)
{
return LZ4_decompress_generic(source, dest, compressedSize, maxDecompressedSize,
@@ -2153,7 +2153,7 @@ int LZ4_decompress_safe(const char* source, char* dest, int compressedSize, int
(BYTE*)dest, NULL, 0);
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_safe_partial(const char* src, char* dst, int compressedSize, int targetOutputSize, int dstCapacity)
{
dstCapacity = MIN(targetOutputSize, dstCapacity);
@@ -2162,7 +2162,7 @@ int LZ4_decompress_safe_partial(const char* src, char* dst, int compressedSize,
noDict, (BYTE*)dst, NULL, 0);
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_fast(const char* source, char* dest, int originalSize)
{
return LZ4_decompress_generic(source, dest, 0, originalSize,
@@ -2172,7 +2172,7 @@ int LZ4_decompress_fast(const char* source, char* dest, int originalSize)
/*===== Instantiate a few more decoding cases, used more than once. =====*/
-LZ4_FORCE_O2_GCC_PPC64LE /* Exported, an obsolete API function. */
+LZ4_FORCE_O2 /* Exported, an obsolete API function. */
int LZ4_decompress_safe_withPrefix64k(const char* source, char* dest, int compressedSize, int maxOutputSize)
{
return LZ4_decompress_generic(source, dest, compressedSize, maxOutputSize,
@@ -2188,7 +2188,7 @@ int LZ4_decompress_fast_withPrefix64k(const char* source, char* dest, int origin
return LZ4_decompress_fast(source, dest, originalSize);
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
static int LZ4_decompress_safe_withSmallPrefix(const char* source, char* dest, int compressedSize, int maxOutputSize,
size_t prefixSize)
{
@@ -2197,7 +2197,7 @@ static int LZ4_decompress_safe_withSmallPrefix(const char* source, char* dest, i
(BYTE*)dest-prefixSize, NULL, 0);
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_safe_forceExtDict(const char* source, char* dest,
int compressedSize, int maxOutputSize,
const void* dictStart, size_t dictSize)
@@ -2207,7 +2207,7 @@ int LZ4_decompress_safe_forceExtDict(const char* source, char* dest,
(BYTE*)dest, (const BYTE*)dictStart, dictSize);
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
static int LZ4_decompress_fast_extDict(const char* source, char* dest, int originalSize,
const void* dictStart, size_t dictSize)
{
@@ -2296,7 +2296,7 @@ int LZ4_decoderRingBufferSize(int maxBlockSize)
If it's not possible, save the relevant part of decoded data into a safe buffer,
and indicate where it stands using LZ4_setStreamDecode()
*/
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* source, char* dest, int compressedSize, int maxOutputSize)
{
LZ4_streamDecode_t_internal* lz4sd = &LZ4_streamDecode->internal_donotuse;
@@ -2336,7 +2336,7 @@ int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
return result;
}
-LZ4_FORCE_O2_GCC_PPC64LE
+LZ4_FORCE_O2
int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* source, char* dest, int originalSize)
{
LZ4_streamDecode_t_internal* lz4sd = &LZ4_streamDecode->internal_donotuse;
diff --git a/lib/lz4hc.c b/lib/lz4hc.c
index 080816e..a126c83 100644
--- a/lib/lz4hc.c
+++ b/lib/lz4hc.c
@@ -158,7 +158,7 @@ int LZ4HC_countBack(const BYTE* const ip, const BYTE* const match,
#endif
-static U32 LZ4HC_rotatePattern(size_t const rotate, U32 const pattern)
+LZ4_FORCE_INLINE U32 LZ4HC_rotatePattern(size_t const rotate, U32 const pattern)
{
size_t const bitsToRotate = (rotate & (sizeof(pattern) - 1)) << 3;
if (bitsToRotate == 0)
@@ -223,7 +223,7 @@ LZ4HC_reverseCountPattern(const BYTE* ip, const BYTE* const iLow, U32 pattern)
* 4 byte MINMATCH would overflow.
* @returns true if the match index is okay.
*/
-static int LZ4HC_protectDictEnd(U32 const dictLimit, U32 const matchIndex)
+LZ4_FORCE_INLINE int LZ4HC_protectDictEnd(U32 const dictLimit, U32 const matchIndex)
{
return ((U32)((dictLimit - 1) - matchIndex) >= 3);
}