diff options
| author | Jia Tan <jiat0218@gmail.com> | 2024-01-22 21:36:09 +0800 |
|---|---|---|
| committer | Jia Tan <jiat0218@gmail.com> | 2024-02-01 20:09:11 +0800 |
| commit | 1940f0ec28f08c0ac72c1413d9706fb82eabe6ad (patch) | |
| tree | 2df2dc8e2e9ea84e725b6b67b906f668c70eed64 | |
| parent | liblzma: Rename crc32_aarch64.h to crc32_arm64.h. (diff) | |
| download | xz-1940f0ec28f08c0ac72c1413d9706fb82eabe6ad.tar.xz | |
liblzma: Omit CRC tables when not needed with ARM64 optimizations.
This is similar to the existing x86-64 CLMUL conditions to omit the
tables. They were slightly refactored to improve readability.
| -rw-r--r-- | src/liblzma/check/crc32_table.c | 18 | ||||
| -rw-r--r-- | src/liblzma/check/crc64_table.c | 7 | ||||
| -rw-r--r-- | src/liblzma/check/crc_common.h | 5 |
3 files changed, 25 insertions, 5 deletions
diff --git a/src/liblzma/check/crc32_table.c b/src/liblzma/check/crc32_table.c index e53b63c9..41a80dcb 100644 --- a/src/liblzma/check/crc32_table.c +++ b/src/liblzma/check/crc32_table.c @@ -13,11 +13,23 @@ #include "common.h" -// FIXME: Compared to crc32_fast.c this has to check for __x86_64__ too +// FIXME: Compared to crc_common.h this has to check for __x86_64__ too // so that in 32-bit builds crc32_x86.S won't break due to a missing table. -#if !defined(HAVE_ENCODERS) && ((defined(__x86_64__) && defined(__SSSE3__) \ +#if (defined(__x86_64__) && defined(__SSSE3__) \ && defined(__SSE4_1__) && defined(__PCLMUL__)) \ - || (defined(__e2k__) && __iset__ >= 6)) + || (defined(__e2k__) && __iset__ >= 6) +# define X86_CLMUL_NO_TABLE 1 +#endif + +#if defined(HAVE_ARM64_CRC32) \ + && !defined(WORDS_BIGENDIAN) \ + && defined(__ARM_FEATURE_CRC32) +# define ARM64_CRC32_NO_TABLE 1 +#endif + + +#if !defined(HAVE_ENCODERS) && (defined(X86_CLMUL_NO_TABLE) \ + || defined(ARM64_CRC32_NO_TABLE_)) // No table needed. Use a typedef to avoid an empty translation unit. typedef void lzma_crc32_dummy; diff --git a/src/liblzma/check/crc64_table.c b/src/liblzma/check/crc64_table.c index 688e527b..0ae9dda8 100644 --- a/src/liblzma/check/crc64_table.c +++ b/src/liblzma/check/crc64_table.c @@ -13,11 +13,16 @@ #include "common.h" -// FIXME: Compared to crc64_fast.c this has to check for __x86_64__ too +// FIXME: Compared to crc_common.h this has to check for __x86_64__ too // so that in 32-bit builds crc64_x86.S won't break due to a missing table. #if (defined(__x86_64__) && defined(__SSSE3__) \ && defined(__SSE4_1__) && defined(__PCLMUL__)) \ || (defined(__e2k__) && __iset__ >= 6) +# define X86_CLMUL_NO_TABLE 1 +#endif + + +#ifdef X86_CLMUL_NO_TABLE // No table needed. Use a typedef to avoid an empty translation unit. typedef void lzma_crc64_dummy; diff --git a/src/liblzma/check/crc_common.h b/src/liblzma/check/crc_common.h index 35f60d95..123d7db0 100644 --- a/src/liblzma/check/crc_common.h +++ b/src/liblzma/check/crc_common.h @@ -75,6 +75,9 @@ // ARM64 CRC32 instruction is only useful for CRC32. Currently, only // little endian is supported since we were unable to test on a big // endian machine. +// +// NOTE: Keep this and the next check in sync with the macro +// ARM64_CRC32_NO_TABLE in crc32_table.c #if defined(HAVE_ARM64_CRC32) && !defined(WORDS_BIGENDIAN) // Allow ARM64 CRC32 instruction without a runtime check if // __ARM_FEATURE_CRC32 is defined. GCC and Clang only define this if the @@ -94,7 +97,7 @@ // generic version can be omitted. Note that this doesn't work with MSVC // as I don't know how to detect the features here. // -// NOTE: Keep this this in sync with crc32_table.c. +// NOTE: Keep this in sync with the CLMUL_NO_TABLE macro in crc32_table.c. # if (defined(__SSSE3__) && defined(__SSE4_1__) && defined(__PCLMUL__)) \ || (defined(__e2k__) && __iset__ >= 6) # define CRC32_ARCH_OPTIMIZED 1 |