[llvm] r244915 - [ARM] Allow vmin/vmax of scalars to be emitted without UseNEONForFP.

[James Molloy via llvm-commits]

Author: jamesm
Date: Thu Aug 13 12:28:20 2015
New Revision: 244915

URL: http://llvm.org/viewvc/llvm-project?rev=244915&view=rev
Log:
[ARM] Allow vmin/vmax of scalars to be emitted without UseNEONForFP.

This overrides the default to more closely resemble the hand-crafted matching logic in ISelLowering. It makes sense, as there is no VFP equivalent of vmin or vmax, to use them when they're available even if in general VFP ops should be preferred.

This should be NFC.

Modified:
    llvm/trunk/lib/Target/ARM/ARMInstrNEON.td

Modified: llvm/trunk/lib/Target/ARM/ARMInstrNEON.td
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM/ARMInstrNEON.td?rev=244915&r1=244914&r2=244915&view=diff
==============================================================================
--- llvm/trunk/lib/Target/ARM/ARMInstrNEON.td (original)
+++ llvm/trunk/lib/Target/ARM/ARMInstrNEON.td Thu Aug 13 12:28:20 2015
@@ -6338,8 +6338,8 @@ def : N3VSMulOpPat<fmul, fsub, VFMSfd>,
       Requires<[HasVFP4, UseNEONForFP, UseFusedMAC]>;
 def : N2VSPat<fabs, VABSfd>;
 def : N2VSPat<fneg, VNEGfd>;
-def : N3VSPat<fmaxnan, VMAXfd>;
-def : N3VSPat<fminnan, VMINfd>;
+def : N3VSPat<fmaxnan, VMAXfd>, Requires<[HasNEON]>;
+def : N3VSPat<fminnan, VMINfd>, Requires<[HasNEON]>;
 def : NVCVTFIPat<fp_to_sint, VCVTf2sd>;
 def : NVCVTFIPat<fp_to_uint, VCVTf2ud>;
 def : NVCVTIFPat<sint_to_fp, VCVTs2fd>;