diff options
author | Aurelien Jarno <aurelien@aurel32.net> | 2022-10-03 23:16:46 +0200 |
---|---|---|
committer | Aurelien Jarno <aurelien@aurel32.net> | 2022-10-03 23:53:06 +0200 |
commit | 7e1326c43a7ab2468946bd711983907d79ceca3b (patch) | |
tree | d01c929f8203bcdf9101e7ee10ecb3a781fde97f | |
parent | 6260de8867d9634e5b9b430ce321da6a3256e9a5 (diff) | |
download | glibc-7e1326c43a7ab2468946bd711983907d79ceca3b.tar glibc-7e1326c43a7ab2468946bd711983907d79ceca3b.tar.gz glibc-7e1326c43a7ab2468946bd711983907d79ceca3b.tar.bz2 glibc-7e1326c43a7ab2468946bd711983907d79ceca3b.zip |
x86-64: Require BMI2 for AVX2 strncmp implementation
The AVX2 strncmp implementations uses the 'bzhi' instruction, which
belongs to the BMI2 CPU feature.
NB: It also uses the 'tzcnt' BMI1 instruction, but it is executed as BSF
as BSF if the CPU doesn't support TZCNT, and produces the same result
for non-zero input.
Partially fixes: b77b06e0e296 ("x86: Optimize strcmp-avx2.S")
Partially resolves: BZ #29611
Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
(cherry picked from commit fc7de1d9b99ae1676bc626ddca422d7abee0eb48)
-rw-r--r-- | sysdeps/x86_64/multiarch/ifunc-impl-list.c | 7 | ||||
-rw-r--r-- | sysdeps/x86_64/multiarch/strncmp.c | 4 |
2 files changed, 7 insertions, 4 deletions
diff --git a/sysdeps/x86_64/multiarch/ifunc-impl-list.c b/sysdeps/x86_64/multiarch/ifunc-impl-list.c index 136e3b8103..b27ae975e4 100644 --- a/sysdeps/x86_64/multiarch/ifunc-impl-list.c +++ b/sysdeps/x86_64/multiarch/ifunc-impl-list.c @@ -1040,15 +1040,18 @@ __libc_ifunc_impl_list (const char *name, struct libc_ifunc_impl *array, /* Support sysdeps/x86_64/multiarch/strncmp.c. */ IFUNC_IMPL (i, name, strncmp, IFUNC_IMPL_ADD (array, i, strncmp, - CPU_FEATURE_USABLE (AVX2), + (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2)), __strncmp_avx2) IFUNC_IMPL_ADD (array, i, strncmp, (CPU_FEATURE_USABLE (AVX2) + && CPU_FEATURE_USABLE (BMI2) && CPU_FEATURE_USABLE (RTM)), __strncmp_avx2_rtm) IFUNC_IMPL_ADD (array, i, strncmp, (CPU_FEATURE_USABLE (AVX512VL) - && CPU_FEATURE_USABLE (AVX512BW)), + && CPU_FEATURE_USABLE (AVX512BW) + && CPU_FEATURE_USABLE (BMI2)), __strncmp_evex) IFUNC_IMPL_ADD (array, i, strncmp, CPU_FEATURE_USABLE (SSE4_2), __strncmp_sse42) diff --git a/sysdeps/x86_64/multiarch/strncmp.c b/sysdeps/x86_64/multiarch/strncmp.c index fca74199d8..4241da5bdf 100644 --- a/sysdeps/x86_64/multiarch/strncmp.c +++ b/sysdeps/x86_64/multiarch/strncmp.c @@ -39,11 +39,11 @@ IFUNC_SELECTOR (void) const struct cpu_features* cpu_features = __get_cpu_features (); if (CPU_FEATURE_USABLE_P (cpu_features, AVX2) + && CPU_FEATURE_USABLE_P (cpu_features, BMI2) && CPU_FEATURES_ARCH_P (cpu_features, AVX_Fast_Unaligned_Load)) { if (CPU_FEATURE_USABLE_P (cpu_features, AVX512VL) - && CPU_FEATURE_USABLE_P (cpu_features, AVX512BW) - && CPU_FEATURE_USABLE_P (cpu_features, BMI2)) + && CPU_FEATURE_USABLE_P (cpu_features, AVX512BW)) return OPTIMIZE (evex); if (CPU_FEATURE_USABLE_P (cpu_features, RTM)) |