668eaab0c7
* Fri Jul 22 2022 Arjun Shankar <arjun@redhat.com> - 2.34-40
- Sync with upstream branch release/2.34/master,
commit b2f32e746492615a6eb3e66fac1e766e32e8deb1:
- malloc: Simplify implementation of __malloc_assert
- Update syscall-names.list for Linux 5.18
- x86: Add missing IS_IN (libc) check to strncmp-sse4_2.S
- x86: Move mem{p}{mov|cpy}_{chk_}erms to its own file
- x86: Move and slightly improve memset_erms
- x86: Add definition for __wmemset_chk AVX2 RTM in ifunc impl list
- x86: Put wcs{n}len-sse4.1 in the sse4.1 text section
- x86: Align entry for memrchr to 64-bytes.
- x86: Add BMI1/BMI2 checks for ISA_V3 check
- x86: Cleanup bounds checking in large memcpy case
- x86: Add bounds `x86_non_temporal_threshold`
- x86: Add sse42 implementation to strcmp's ifunc
- x86: Fix misordered logic for setting `rep_movsb_stop_threshold`
- x86: Align varshift table to 32-bytes
- x86: ZERO_UPPER_VEC_REGISTERS_RETURN_XTEST expect no transactions
- x86: Shrink code size of memchr-evex.S
- x86: Shrink code size of memchr-avx2.S
- x86: Optimize memrchr-avx2.S
- x86: Optimize memrchr-evex.S
- x86: Optimize memrchr-sse2.S
- x86: Add COND_VZEROUPPER that can replace vzeroupper if no `ret`
- x86: Create header for VEC classes in x86 strings library
- x86_64: Add strstr function with 512-bit EVEX
- x86-64: Ignore r_addend for R_X86_64_GLOB_DAT/R_X86_64_JUMP_SLOT
- x86_64: Implement evex512 version of strlen, strnlen, wcslen and wcsnlen
- x86_64: Remove bzero optimization
- x86_64: Remove end of line trailing spaces
- nptl: Fix ___pthread_unregister_cancel_restore asynchronous restore
- linux: Fix mq_timereceive check for 32 bit fallback code (BZ 29304)
Resolves: #2109505
39 lines
1.5 KiB
Diff
39 lines
1.5 KiB
Diff
commit 9d50e162eef88e1f870a941b0a973060e984e7ca
|
|
Author: Noah Goldstein <goldstein.w.n@gmail.com>
|
|
Date: Tue Jun 14 15:37:28 2022 -0700
|
|
|
|
x86: Add sse42 implementation to strcmp's ifunc
|
|
|
|
This has been missing since the the ifuncs where added.
|
|
|
|
The performance of SSE4.2 is preferable to to SSE2.
|
|
|
|
Measured on Tigerlake with N = 20 runs.
|
|
Geometric Mean of all benchmarks SSE4.2 / SSE2: 0.906
|
|
|
|
(cherry picked from commit ff439c47173565fbff4f0f78d07b0f14e4a7db05)
|
|
|
|
diff --git a/sysdeps/x86_64/multiarch/strcmp.c b/sysdeps/x86_64/multiarch/strcmp.c
|
|
index 7c2901bf44456259..b457fb4c150e4407 100644
|
|
--- a/sysdeps/x86_64/multiarch/strcmp.c
|
|
+++ b/sysdeps/x86_64/multiarch/strcmp.c
|
|
@@ -29,6 +29,7 @@
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (sse2) attribute_hidden;
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (sse2_unaligned) attribute_hidden;
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (ssse3) attribute_hidden;
|
|
+extern __typeof (REDIRECT_NAME) OPTIMIZE (sse42) attribute_hidden;
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2) attribute_hidden;
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (avx2_rtm) attribute_hidden;
|
|
extern __typeof (REDIRECT_NAME) OPTIMIZE (evex) attribute_hidden;
|
|
@@ -53,6 +54,10 @@ IFUNC_SELECTOR (void)
|
|
return OPTIMIZE (avx2);
|
|
}
|
|
|
|
+ if (CPU_FEATURE_USABLE_P (cpu_features, SSE4_2)
|
|
+ && !CPU_FEATURES_ARCH_P (cpu_features, Slow_SSE4_2))
|
|
+ return OPTIMIZE (sse42);
|
|
+
|
|
if (CPU_FEATURES_ARCH_P (cpu_features, Fast_Unaligned_Load))
|
|
return OPTIMIZE (sse2_unaligned);
|
|
|