[C-API] Add LLVMGetHostCPU{Name,Features}.

[anarazel]

Summary

Without these functions it's hard to create a TargetMachine for Orc JIT that creates efficient native code.

It's not sufficient to just expose LLVMGetHostCPUName(), because for some CPUs there's fewer features
actually available than the CPU name indicates (e.g. AVX might be missing on some CPUs identified as skylake).

[anarazel]

Ping? This is a trivial change and makes it much easier to use Orc via the C API, as otherwise suboptimal code is generated. If anybody has suggestions for further reviewers I'd also be happy.

[whitequark]

LGTM. Ideally we would have a test for this code but given that it involves host CPU and features this seems quite tricky and not worth the benefit.

[anarazel]

Ideally we would have a test for this code but given that it involves host CPU and features this seems quite tricky and not worth the benefit.

Yea, I was wondering about that, but the only test that seemed possible with reasonable would be to just call the functions and throw the results away, which didn't seem like it'd add much.

Thanks!

llvm/trunk/include/llvm-c/TargetMachine.h

 LLVMBool LLVMTargetMachineEmitToMemoryBuffer(LLVMTargetMachineRef T, LLVMModuleRef M,
   LLVMCodeGenFileType codegen, char** ErrorMessage, LLVMMemoryBufferRef *OutMemBuf);
 
 /*===-- Triple ------------------------------------------------------------===*/
 /** Get a triple for the host machine as a string. The result needs to be
   disposed with LLVMDisposeMessage. */
 char* LLVMGetDefaultTargetTriple(void);
 
+/** Get the host CPU as a string. The result needs to be disposed with
+  LLVMDisposeMessage. */
+char* LLVMGetHostCPUName(void);
+
+/** Get the host CPU's features as a string. The result needs to be disposed
+  with LLVMDisposeMessage. */
+char* LLVMGetHostCPUFeatures(void);
+
 /** Adds the target-specific analysis passes to the pass manager. */
 void LLVMAddAnalysisPasses(LLVMTargetMachineRef T, LLVMPassManagerRef PM);
 
 #ifdef __cplusplus
 }
 #endif
 
 #endif

llvm/trunk/lib/Target/TargetMachineC.cpp

 
 #include "llvm-c/Core.h"
 #include "llvm-c/Target.h"
 #include "llvm-c/TargetMachine.h"
 #include "llvm/Analysis/TargetTransformInfo.h"
 #include "llvm/IR/DataLayout.h"
 #include "llvm/IR/LegacyPassManager.h"
 #include "llvm/IR/Module.h"
+#include "llvm/MC/SubtargetFeature.h"
 #include "llvm/Support/FileSystem.h"
 #include "llvm/Support/FormattedStream.h"
 #include "llvm/Support/Host.h"
 #include "llvm/Support/TargetRegistry.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Target/CodeGenCWrappers.h"
 #include "llvm/Target/TargetMachine.h"
 #include <cassert>
*OutMemBuf =
       LLVMCreateMemoryBufferWithMemoryRangeCopy(Data.data(), Data.size(), "");
   return Result;
 }
 
 char *LLVMGetDefaultTargetTriple(void) {
   return strdup(sys::getDefaultTargetTriple().c_str());
 }
 
+char *LLVMGetHostCPUName(void) {
+  return strdup(sys::getHostCPUName().data());
+}
+
+char *LLVMGetHostCPUFeatures(void) {
+  SubtargetFeatures Features;
+  StringMap<bool> HostFeatures;
+
+  if (sys::getHostCPUFeatures(HostFeatures))
+    for (auto &F : HostFeatures)
+      Features.AddFeature(F.first(), F.second);
+
+  return strdup(Features.getString().c_str());
+}
+
 void LLVMAddAnalysisPasses(LLVMTargetMachineRef T, LLVMPassManagerRef PM) {
   unwrap(PM)->add(
       createTargetTransformInfoWrapperPass(unwrap(T)->getTargetIRAnalysis()));
 }