[libc-commits] [libc] da5d00a - [libc] Fix builtin definition for memory functions

[Michael Jones via libc-commits]

Author: Michael Jones
Date: 2022-11-16T13:03:39-08:00
New Revision: da5d00ad0cf4d45e38287280c670b08e84e27411

URL: https://github.com/llvm/llvm-project/commit/da5d00ad0cf4d45e38287280c670b08e84e27411
DIFF: https://github.com/llvm/llvm-project/commit/da5d00ad0cf4d45e38287280c670b08e84e27411.diff

LOG: [libc] Fix builtin definition for memory functions

The memory functions are highly performance sensitive and use builtins
where possible, but also need to define those functions names when they
don't exist to avoid compilation errors. Previously all those
redefinitions were behind the SSE2 flag for x86, which caused errors on
CPUs that supported SSE2 but not AVX512. This patch splits the various
CPU extensions out to avoid errors on such CPUs.

Reviewed By: gchatelet

Differential Revision: https://reviews.llvm.org/D137868

Added: 
    

Modified: 
    libc/cmake/modules/LLVMLibCCheckCpuFeatures.cmake
    libc/src/string/CMakeLists.txt
    libc/src/string/memory_utils/op_x86.h

Removed: 
    


################################################################################
diff  --git a/libc/cmake/modules/LLVMLibCCheckCpuFeatures.cmake b/libc/cmake/modules/LLVMLibCCheckCpuFeatures.cmake
index 358d07480ab80..fdd477c1155ec 100644
--- a/libc/cmake/modules/LLVMLibCCheckCpuFeatures.cmake
+++ b/libc/cmake/modules/LLVMLibCCheckCpuFeatures.cmake
@@ -6,7 +6,7 @@
 set(ALL_CPU_FEATURES "")
 
 if(${LIBC_TARGET_ARCHITECTURE_IS_X86})
-  set(ALL_CPU_FEATURES SSE2 SSE4_2 AVX2 AVX512F FMA)
+  set(ALL_CPU_FEATURES SSE2 SSE4_2 AVX2 AVX512F AVX512BW FMA)
   set(LIBC_COMPILE_OPTIONS_NATIVE -march=native)
 elseif(${LIBC_TARGET_ARCHITECTURE_IS_AARCH64})
   set(LIBC_COMPILE_OPTIONS_NATIVE -mcpu=native)

diff  --git a/libc/src/string/CMakeLists.txt b/libc/src/string/CMakeLists.txt
index e300a036655ab..9e8287fe8b938 100644
--- a/libc/src/string/CMakeLists.txt
+++ b/libc/src/string/CMakeLists.txt
@@ -355,7 +355,7 @@ if(${LIBC_TARGET_ARCHITECTURE_IS_X86})
   add_bcmp(bcmp_x86_64_opt_sse2   COMPILE_OPTIONS -march=k8             REQUIRE SSE2)
   add_bcmp(bcmp_x86_64_opt_sse4   COMPILE_OPTIONS -march=nehalem        REQUIRE SSE4_2)
   add_bcmp(bcmp_x86_64_opt_avx2   COMPILE_OPTIONS -march=haswell        REQUIRE AVX2)
-  add_bcmp(bcmp_x86_64_opt_avx512 COMPILE_OPTIONS -march=skylake-avx512 REQUIRE AVX512F)
+  add_bcmp(bcmp_x86_64_opt_avx512 COMPILE_OPTIONS -march=skylake-avx512 REQUIRE AVX512BW)
   add_bcmp(bcmp_opt_host          COMPILE_OPTIONS ${LIBC_COMPILE_OPTIONS_NATIVE})
   add_bcmp(bcmp)
 else()
@@ -409,7 +409,7 @@ if(${LIBC_TARGET_ARCHITECTURE_IS_X86})
   add_memcmp(memcmp_x86_64_opt_sse2   COMPILE_OPTIONS -march=k8             REQUIRE SSE2)
   add_memcmp(memcmp_x86_64_opt_sse4   COMPILE_OPTIONS -march=nehalem        REQUIRE SSE4_2)
   add_memcmp(memcmp_x86_64_opt_avx2   COMPILE_OPTIONS -march=haswell        REQUIRE AVX2)
-  add_memcmp(memcmp_x86_64_opt_avx512 COMPILE_OPTIONS -march=skylake-avx512 REQUIRE AVX512F)
+  add_memcmp(memcmp_x86_64_opt_avx512 COMPILE_OPTIONS -march=skylake-avx512 REQUIRE AVX512BW)
   add_memcmp(memcmp_opt_host          COMPILE_OPTIONS ${LIBC_COMPILE_OPTIONS_NATIVE})
   add_memcmp(memcmp)
 elseif(${LIBC_TARGET_ARCHITECTURE_IS_AARCH64})

diff  --git a/libc/src/string/memory_utils/op_x86.h b/libc/src/string/memory_utils/op_x86.h
index 8e6432233ca31..a729cf28bbf4b 100644
--- a/libc/src/string/memory_utils/op_x86.h
+++ b/libc/src/string/memory_utils/op_x86.h
@@ -20,15 +20,22 @@
 #include "src/string/memory_utils/op_builtin.h"
 #include "src/string/memory_utils/op_generic.h"
 
-#ifdef __SSE2__
+#if defined(__AVX512BW__) || defined(__AVX512F__) || defined(__AVX2__) ||      \
+    defined(__SSE2__)
 #include <immintrin.h>
-#else
+#endif
+
 // Define fake functions to prevent the compiler from failing on undefined
-// functions in case SSE2 is not present.
+// functions in case the CPU extension is not present.
+#ifndef __AVX512BW__
 #define _mm512_cmpneq_epi8_mask(A, B) 0
-#define _mm_movemask_epi8(A) 0
+#endif
+#ifndef __AVX2__
 #define _mm256_movemask_epi8(A) 0
-#endif //  __SSE2__
+#endif
+#ifndef __SSE2__
+#define _mm_movemask_epi8(A) 0
+#endif
 
 namespace __llvm_libc::x86 {