[PATCH] D70214: [X86] Add custom type legalization and lowering for scalar STRICT_FP_TO_SINT/UINT

[Craig Topper via Phabricator via llvm-commits]

craig.topper updated this revision to Diff 229579.
craig.topper added a comment.

Use strict fp fsub in the expansion of unsigned conversion to i64 on 32-bit targets without sse support for the floating point type. This technically can still raise a spurious inexact exception, but the same thing happens with the default expansion in LegalizeDAG we use on i64 targets.

Use a different form of getNode to avoid explicitly calling getVTList.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70214/new/

https://reviews.llvm.org/D70214

Files:
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/lib/Target/X86/X86ISelLowering.h
  llvm/test/CodeGen/X86/fp-intrinsics-x86_64.ll
  llvm/test/CodeGen/X86/fp-intrinsics.ll
  llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D70214.229579.patch
Type: text/x-patch
Size: 58584 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20191115/8797dc37/attachment-0001.bin>