[PATCH] D59657: [LangRef] Clarify codegen expectations for intrinsics with fp/integer-only overloads

[Amara Emerson via Phabricator via llvm-commits]

aemerson created this revision.
aemerson added reviewers: eli.friedman, arsenm, qcolombet.
aemerson added a project: LLVM.
Herald added a subscriber: wdng.

[LangRef] Clarify codegen expectations for intrinsics with fp/integer-only overloads.

This change is a result of discussions on list: "GlobalISel: Ambiguous intrinsic semantics problem"


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D59657

Files:
  llvm/docs/LangRef.rst


Index: llvm/docs/LangRef.rst
===================================================================
--- llvm/docs/LangRef.rst
+++ llvm/docs/LangRef.rst
@@ -10394,6 +10394,12 @@
 type is matched against the return type, it does not require its own
 name suffix.
 
+Intrinsic overloads based solely the distinction between whether the
+types are integer or floating point should not be relied upon for
+correct code generation. In such cases, the recommended approach when
+defining intrinsics is to create separate integer and FP intrinsics
+rather than rely on overloading.
+
 To learn how to add an intrinsic function, please see the `Extending
 LLVM Guide <ExtendingLLVM.html>`_.
 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D59657.191749.patch
Type: text/x-patch
Size: 681 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190321/087749bb/attachment.bin>