diff options
| author | Jia Tan <jiat0218@gmail.com> | 2023-10-18 19:57:10 +0800 |
|---|---|---|
| committer | Jia Tan <jiat0218@gmail.com> | 2023-10-18 23:54:41 +0800 |
| commit | 1c8884f0af28b3a4690bb573cdf3240a8ec73416 (patch) | |
| tree | 324c32b0dc4d7cd3217a3d394965772204b4ed80 | |
| parent | liblzma: CRC_USE_GENERIC_FOR_SMALL_INPUTS cannot be used with ifunc. (diff) | |
| download | xz-1c8884f0af28b3a4690bb573cdf3240a8ec73416.tar.xz | |
liblzma: Set the MSVC optimization fix to only cover lzma_crc64_clmul().
After testing a 32-bit Release build on MSVC, only lzma_crc64_clmul()
has the bug. crc_simd_body() and lzma_crc32_clmul() do not need the
optimizations disabled.
| -rw-r--r-- | src/liblzma/check/crc_clmul.c | 30 |
1 files changed, 15 insertions, 15 deletions
diff --git a/src/liblzma/check/crc_clmul.c b/src/liblzma/check/crc_clmul.c index 404adb82..b90f43e3 100644 --- a/src/liblzma/check/crc_clmul.c +++ b/src/liblzma/check/crc_clmul.c @@ -42,21 +42,6 @@ MASK_H(in, mask, high) -// MSVC (VS2015 - VS2022) produces bad 32-bit x86 code from the CLMUL CRC -// code when optimizations are enabled (release build). According to the bug -// report, the ebx register is corrupted and the calculated result is wrong. -// Trying to workaround the problem with "__asm mov ebx, ebx" didn't help. -// The following pragma works and performance is still good. x86-64 builds -// aren't affected by this problem. -// -// NOTE: Another pragma after lzma_crc64_clmul() restores the optimizations. -// If the #if condition here is updated, the other one must be updated too. -#if defined(_MSC_VER) && !defined(__INTEL_COMPILER) && !defined(__clang__) \ - && defined(_M_IX86) -# pragma optimize("g", off) -#endif - - #if (defined(__GNUC__) || defined(__clang__)) && !defined(__EDG__) __attribute__((__target__("ssse3,sse4.1,pclmul"))) #endif @@ -314,6 +299,21 @@ calc_hi(uint64_t poly, uint64_t a) #ifdef HAVE_CHECK_CRC64 +// MSVC (VS2015 - VS2022) produces bad 32-bit x86 code from the CLMUL CRC +// code when optimizations are enabled (release build). According to the bug +// report, the ebx register is corrupted and the calculated result is wrong. +// Trying to workaround the problem with "__asm mov ebx, ebx" didn't help. +// The following pragma works and performance is still good. x86-64 builds +// and CRC32 CLMUL aren't affected by this problem. The problem does not +// happen in crc_simd_body() either (which is shared with CRC32 CLMUL anyway). +// +// NOTE: Another pragma after lzma_crc64_clmul() restores the optimizations. +// If the #if condition here is updated, the other one must be updated too. +#if defined(_MSC_VER) && !defined(__INTEL_COMPILER) && !defined(__clang__) \ + && defined(_M_IX86) +# pragma optimize("g", off) +#endif + #if (defined(__GNUC__) || defined(__clang__)) && !defined(__EDG__) __attribute__((__target__("ssse3,sse4.1,pclmul"))) #endif |