+#if defined(i386) || defined(__i386__) || defined(__x86__) || defined(_M_IX86)\
+ || defined(__x86_64__) || defined(_M_AMD64) || defined (_M_X64)
+bool sys::getHostCPUFeatures(StringMap<bool> &Features) {
+ unsigned EAX = 0, EBX = 0, ECX = 0, EDX = 0;
+ unsigned MaxLevel;
+ union {
+ unsigned u[3];
+ char c[12];
+ } text;
+
+ if (GetX86CpuIDAndInfo(0, &MaxLevel, text.u+0, text.u+2, text.u+1) ||
+ MaxLevel < 1)
+ return false;
+
+ GetX86CpuIDAndInfo(1, &EAX, &EBX, &ECX, &EDX);
+
+ Features["cmov"] = (EDX >> 15) & 1;
+ Features["mmx"] = (EDX >> 23) & 1;
+ Features["sse"] = (EDX >> 25) & 1;
+ Features["sse2"] = (EDX >> 26) & 1;
+ Features["sse3"] = (ECX >> 0) & 1;
+ Features["ssse3"] = (ECX >> 9) & 1;
+ Features["sse4.1"] = (ECX >> 19) & 1;
+ Features["sse4.2"] = (ECX >> 20) & 1;
+
+ Features["pclmul"] = (ECX >> 1) & 1;
+ Features["cx16"] = (ECX >> 13) & 1;
+ Features["movbe"] = (ECX >> 22) & 1;
+ Features["popcnt"] = (ECX >> 23) & 1;
+ Features["aes"] = (ECX >> 25) & 1;
+ Features["rdrnd"] = (ECX >> 30) & 1;
+
+ // If CPUID indicates support for XSAVE, XRESTORE and AVX, and XGETBV
+ // indicates that the AVX registers will be saved and restored on context
+ // switch, then we have full AVX support.
+ bool HasAVX = ((ECX >> 27) & 1) && ((ECX >> 28) & 1) &&
+ !GetX86XCR0(&EAX, &EDX) && ((EAX & 0x6) == 0x6);
+ Features["avx"] = HasAVX;
+ Features["fma"] = HasAVX && (ECX >> 12) & 1;
+ Features["f16c"] = HasAVX && (ECX >> 29) & 1;
+
+ // AVX512 requires additional context to be saved by the OS.
+ bool HasAVX512Save = HasAVX && ((EAX & 0xe0) == 0xe0);
+
+ unsigned MaxExtLevel;
+ GetX86CpuIDAndInfo(0x80000000, &MaxExtLevel, &EBX, &ECX, &EDX);
+
+ bool HasExtLeaf1 = MaxExtLevel >= 0x80000001 &&
+ !GetX86CpuIDAndInfo(0x80000001, &EAX, &EBX, &ECX, &EDX);
+ Features["lzcnt"] = HasExtLeaf1 && ((ECX >> 5) & 1);
+ Features["sse4a"] = HasExtLeaf1 && ((ECX >> 6) & 1);
+ Features["prfchw"] = HasExtLeaf1 && ((ECX >> 8) & 1);
+ Features["xop"] = HasAVX && HasExtLeaf1 && ((ECX >> 11) & 1);
+ Features["fma4"] = HasAVX && HasExtLeaf1 && ((ECX >> 16) & 1);
+ Features["tbm"] = HasExtLeaf1 && ((ECX >> 21) & 1);
+
+ bool HasLeaf7 = MaxLevel >= 7 &&
+ !GetX86CpuIDAndInfoEx(0x7, 0x0, &EAX, &EBX, &ECX, &EDX);
+
+ // AVX2 is only supported if we have the OS save support from AVX.
+ Features["avx2"] = HasAVX && HasLeaf7 && (EBX >> 5) & 1;
+
+ Features["fsgsbase"] = HasLeaf7 && ((EBX >> 0) & 1);
+ Features["bmi"] = HasLeaf7 && ((EBX >> 3) & 1);
+ Features["hle"] = HasLeaf7 && ((EBX >> 4) & 1);
+ Features["bmi2"] = HasLeaf7 && ((EBX >> 8) & 1);
+ Features["rtm"] = HasLeaf7 && ((EBX >> 11) & 1);
+ Features["rdseed"] = HasLeaf7 && ((EBX >> 18) & 1);
+ Features["adx"] = HasLeaf7 && ((EBX >> 19) & 1);
+ Features["sha"] = HasLeaf7 && ((EBX >> 29) & 1);
+
+ // AVX512 is only supported if the OS supports the context save for it.
+ Features["avx512f"] = HasLeaf7 && ((EBX >> 16) & 1) && HasAVX512Save;
+ Features["avx512dq"] = HasLeaf7 && ((EBX >> 17) & 1) && HasAVX512Save;
+ Features["avx512pf"] = HasLeaf7 && ((EBX >> 26) & 1) && HasAVX512Save;
+ Features["avx512er"] = HasLeaf7 && ((EBX >> 27) & 1) && HasAVX512Save;
+ Features["avx512cd"] = HasLeaf7 && ((EBX >> 28) & 1) && HasAVX512Save;
+ Features["avx512bw"] = HasLeaf7 && ((EBX >> 30) & 1) && HasAVX512Save;
+ Features["avx512vl"] = HasLeaf7 && ((EBX >> 31) & 1) && HasAVX512Save;
+
+ return true;
+}
+#elif defined(__linux__) && (defined(__arm__) || defined(__aarch64__))
+bool sys::getHostCPUFeatures(StringMap<bool> &Features) {
+ // Read 1024 bytes from /proc/cpuinfo, which should contain the Features line
+ // in all cases.
+ char buffer[1024];
+ ssize_t CPUInfoSize = readCpuInfo(buffer, sizeof(buffer));
+ if (CPUInfoSize == -1)
+ return false;
+
+ StringRef Str(buffer, CPUInfoSize);
+
+ SmallVector<StringRef, 32> Lines;
+ Str.split(Lines, "\n");
+
+ SmallVector<StringRef, 32> CPUFeatures;
+
+ // Look for the CPU features.
+ for (unsigned I = 0, E = Lines.size(); I != E; ++I)
+ if (Lines[I].startswith("Features")) {
+ Lines[I].split(CPUFeatures, " ");
+ break;
+ }
+
+#if defined(__aarch64__)
+ // Keep track of which crypto features we have seen
+ enum {
+ CAP_AES = 0x1,
+ CAP_PMULL = 0x2,
+ CAP_SHA1 = 0x4,
+ CAP_SHA2 = 0x8
+ };
+ uint32_t crypto = 0;
+#endif
+
+ for (unsigned I = 0, E = CPUFeatures.size(); I != E; ++I) {
+ StringRef LLVMFeatureStr = StringSwitch<StringRef>(CPUFeatures[I])
+#if defined(__aarch64__)
+ .Case("asimd", "neon")
+ .Case("fp", "fp-armv8")
+ .Case("crc32", "crc")
+#else
+ .Case("half", "fp16")
+ .Case("neon", "neon")
+ .Case("vfpv3", "vfp3")
+ .Case("vfpv3d16", "d16")
+ .Case("vfpv4", "vfp4")
+ .Case("idiva", "hwdiv-arm")
+ .Case("idivt", "hwdiv")
+#endif
+ .Default("");
+
+#if defined(__aarch64__)
+ // We need to check crypto separately since we need all of the crypto
+ // extensions to enable the subtarget feature
+ if (CPUFeatures[I] == "aes")
+ crypto |= CAP_AES;
+ else if (CPUFeatures[I] == "pmull")
+ crypto |= CAP_PMULL;
+ else if (CPUFeatures[I] == "sha1")
+ crypto |= CAP_SHA1;
+ else if (CPUFeatures[I] == "sha2")
+ crypto |= CAP_SHA2;
+#endif
+
+ if (LLVMFeatureStr != "")
+ Features[LLVMFeatureStr] = true;
+ }
+
+#if defined(__aarch64__)
+ // If we have all crypto bits we can add the feature
+ if (crypto == (CAP_AES | CAP_PMULL | CAP_SHA1 | CAP_SHA2))
+ Features["crypto"] = true;
+#endif
+
+ return true;
+}
+#else