[llvm] r334904 - [NVPTX] Ignore target-cpu and -features for inlining

[Jonas Hahnfeld via llvm-commits]

Author: hahnfeld
Date: Sun Jun 17 02:55:20 2018
New Revision: 334904

URL: http://llvm.org/viewvc/llvm-project?rev=334904&view=rev
Log:
[NVPTX] Ignore target-cpu and -features for inlining

We don't want to prevent inlining because of target-cpu and -features
attributes that were added to newer versions of LLVM/Clang: There are
no incompatible functions in PTX, ptxas will throw errors in such cases.

Differential Revision: https://reviews.llvm.org/D47691

Modified:
    llvm/trunk/lib/Target/NVPTX/NVPTXTargetTransformInfo.h

Modified: llvm/trunk/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/NVPTX/NVPTXTargetTransformInfo.h?rev=334904&r1=334903&r2=334904&view=diff
==============================================================================
--- llvm/trunk/lib/Target/NVPTX/NVPTXTargetTransformInfo.h (original)
+++ llvm/trunk/lib/Target/NVPTX/NVPTXTargetTransformInfo.h Sun Jun 17 02:55:20 2018
@@ -61,6 +61,14 @@ public:
   unsigned getRegisterBitWidth(bool Vector) const { return 32; }
   unsigned getMinVectorRegisterBitWidth() const { return 32; }
 
+  // We don't want to prevent inlining because of target-cpu and -features
+  // attributes that were added to newer versions of LLVM/Clang: There are
+  // no incompatible functions in PTX, ptxas will throw errors in such cases.
+  bool areInlineCompatible(const Function *Caller,
+                           const Function *Callee) const {
+    return true;
+  }
+
   // Increase the inlining cost threshold by a factor of 5, reflecting that
   // calls are particularly expensive in NVPTX.
   unsigned getInliningThresholdMultiplier() { return 5; }