[compiler-rt] [compiler-rt] Add cpu model init for Windows. (PR #111961)

[Alexandros Lamprineas via llvm-commits]

================
@@ -0,0 +1,98 @@
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+#include <processthreadsapi.h>
+#include <stdint.h>
+
+#ifndef PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE 43 
+#endif
+#ifndef PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE 44
+#endif
+#ifndef PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE 45
+#endif
+#ifndef PF_ARM_SVE_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_INSTRUCTIONS_AVAILABLE 46
+#endif
+#ifndef PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE 47
+#endif
+#ifndef PF_ARM_SVE_AES_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_AES_INSTRUCTIONS_AVAILABLE 49
+#endif
+#ifndef PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE 50
+#endif
+#ifndef PF_ARM_SVE_BITPERM_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_BITPERM_INSTRUCTIONS_AVAILABLE 51
+#endif
+#ifndef PF_ARM_SVE_BF16_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_BF16_INSTRUCTIONS_AVAILABLE 52
+#endif
+#ifndef PF_ARM_SVE_EBF16_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_EBF16_INSTRUCTIONS_AVAILABLE 53
+#endif
+#ifndef PF_ARM_SVE_SHA3_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_SHA3_INSTRUCTIONS_AVAILABLE 55
+#endif
+#ifndef PF_ARM_SVE_SM4_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_SM4_INSTRUCTIONS_AVAILABLE 56
+#endif
+#ifndef PF_ARM_SVE_I8MM_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_I8MM_INSTRUCTIONS_AVAILABLE 57
+#endif
+#ifndef PF_ARM_SVE_F32MM_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_F32MM_INSTRUCTIONS_AVAILABLE 58
+#endif
+#ifndef PF_ARM_SVE_F64MM_INSTRUCTIONS_AVAILABLE
+#define PF_ARM_SVE_F64MM_INSTRUCTIONS_AVAILABLE 59
+#endif
+
+void __init_cpu_features_resolver(unsigned long hwcap,
+                                  const __ifunc_arg_t *arg) {}
+
+void CONSTRUCTOR_ATTRIBUTE __init_cpu_features(void) {
+  if (__atomic_load_n(&__aarch64_cpu_features.features, __ATOMIC_RELAXED))
+    return;
+
+#define setCPUFeature(F) features |= 1ULL << F
+
+  uint64_t features = 0;
+
+  setCPUFeature(FEAT_INIT);
+  setCPUFeature(FEAT_FP);
+
+  // https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-isprocessorfeaturepresent
+  if (IsProcessorFeaturePresent(PF_ARM_V8_CRYPTO_INSTRUCTIONS_AVAILABLE)) {
+    setCPUFeature(FEAT_AES);
+    setCPUFeature(FEAT_SHA2);
+    setCPUFeature(FEAT_PMULL);
+  }
+
+  static const struct ProcessFeatureToFeatMap_t {
+    int WinApiFeature;
+    enum CPUFeatures CPUFeature;
+  } FeatMap[] = {
+      {PF_ARM_V8_CRC32_INSTRUCTIONS_AVAILABLE, FEAT_CRC},
+      {PF_ARM_V81_ATOMIC_INSTRUCTIONS_AVAILABLE, FEAT_LSE},
+      {PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE, FEAT_DOTPROD},
+      {PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE, FEAT_JSCVT},
+      {PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE, FEAT_RCPC},
+      {PF_ARM_SVE_INSTRUCTIONS_AVAILABLE, FEAT_SVE},
+      {PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE, FEAT_SVE2},
+      {PF_ARM_SVE_AES_INSTRUCTIONS_AVAILABLE, FEAT_SVE_AES},
----------------
labrinea wrote:

This should be replaced with {PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE, FEAT_SVE_PMULL128}. Also the PF_ARM_SVE_AES_INSTRUCTIONS_AVAILABLE macro definition may not be needed any more.

https://github.com/llvm/llvm-project/pull/111961