Add fused multiple+add instructions from VFPv4.

[oota-llvm.git] / lib / Target / ARM / ARMInstrNEON.td

diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td
index a2df5bde880321861ff87df3afe22240c3d88edc..f9626fef00be79c544f2c3d638a8a8af41544ba9 100644 (file)
--- a/lib/Target/ARM/ARMInstrNEON.td
+++ b/lib/Target/ARM/ARMInstrNEON.td
@@ -3897,10 +3897,10 @@ defm VMLA     : N3VMulOp_QHS<0, 0, 0b1001, 0, IIC_VMACi16D, IIC_VMACi32D,
                              IIC_VMACi16Q, IIC_VMACi32Q, "vmla", "i", add>;
 def  VMLAfd   : N3VDMulOp<0, 0, 0b00, 0b1101, 1, IIC_VMACD, "vmla", "f32",
                           v2f32, fmul_su, fadd_mlx>,
-                Requires<[HasNEON, UseFPVMLx]>;
+                Requires<[HasNEON, UseFPVMLx, NoNEONVFP4]>;
 def  VMLAfq   : N3VQMulOp<0, 0, 0b00, 0b1101, 1, IIC_VMACQ, "vmla", "f32",
                           v4f32, fmul_su, fadd_mlx>,
-                Requires<[HasNEON, UseFPVMLx]>;
+                Requires<[HasNEON, UseFPVMLx, NoNEONVFP4]>;
 defm VMLAsl   : N3VMulOpSL_HS<0b0000, IIC_VMACi16D, IIC_VMACi32D,
                               IIC_VMACi16Q, IIC_VMACi32Q, "vmla", "i", add>;
 def  VMLAslfd : N3VDMulOpSL<0b10, 0b0001, IIC_VMACD, "vmla", "f32",
@@ -3955,10 +3955,10 @@ defm VMLS     : N3VMulOp_QHS<1, 0, 0b1001, 0, IIC_VMACi16D, IIC_VMACi32D,
                              IIC_VMACi16Q, IIC_VMACi32Q, "vmls", "i", sub>;
 def  VMLSfd   : N3VDMulOp<0, 0, 0b10, 0b1101, 1, IIC_VMACD, "vmls", "f32",
                           v2f32, fmul_su, fsub_mlx>,
-                Requires<[HasNEON, UseFPVMLx]>;
+                Requires<[HasNEON, UseFPVMLx, NoNEONVFP4]>;
 def  VMLSfq   : N3VQMulOp<0, 0, 0b10, 0b1101, 1, IIC_VMACQ, "vmls", "f32",
                           v4f32, fmul_su, fsub_mlx>,
-                Requires<[HasNEON, UseFPVMLx]>;
+                Requires<[HasNEON, UseFPVMLx, NoNEONVFP4]>;
 defm VMLSsl   : N3VMulOpSL_HS<0b0100, IIC_VMACi16D, IIC_VMACi32D,
                               IIC_VMACi16Q, IIC_VMACi32Q, "vmls", "i", sub>;
 def  VMLSslfd : N3VDMulOpSL<0b10, 0b0101, IIC_VMACD, "vmls", "f32",
@@ -4007,6 +4007,24 @@ defm VQDMLSL  : N3VLInt3_HS<0, 1, 0b1011, 0, IIC_VMACi16D, IIC_VMACi32D,
                             "vqdmlsl", "s", int_arm_neon_vqdmlsl>;
 defm VQDMLSLsl: N3VLInt3SL_HS<0, 0b111, "vqdmlsl", "s", int_arm_neon_vqdmlsl>;
 
+
+// Fused Vector Multiply-Accumulate and Fused Multiply-Subtract Operations.
+def  VFMAfd   : N3VDMulOp<0, 0, 0b00, 0b1100, 1, IIC_VFMACD, "vfma", "f32",
+                          v2f32, fmul_su, fadd_mlx>,
+                Requires<[HasNEONVFP4]>;
+
+def  VFMAfq   : N3VQMulOp<0, 0, 0b00, 0b1100, 1, IIC_VFMACQ, "vfma", "f32",
+                          v4f32, fmul_su, fadd_mlx>,
+                Requires<[HasNEONVFP4]>;
+
+//   Fused Vector Multiply Subtract (floating-point)
+def  VFMSfd   : N3VDMulOp<0, 0, 0b10, 0b1100, 1, IIC_VFMACD, "vfms", "f32",
+                          v2f32, fmul_su, fsub_mlx>,
+                Requires<[HasNEONVFP4]>;
+def  VFMSfq   : N3VQMulOp<0, 0, 0b10, 0b1100, 1, IIC_VFMACQ, "vfms", "f32",
+                          v4f32, fmul_su, fsub_mlx>,
+                Requires<[HasNEONVFP4]>;
+
 // Vector Subtract Operations.
 
 //   VSUB     : Vector Subtract (integer and floating-point)
@@ -5358,9 +5376,13 @@ def : N3VSPat<fadd, VADDfd>;
 def : N3VSPat<fsub, VSUBfd>;
 def : N3VSPat<fmul, VMULfd>;
 def : N3VSMulOpPat<fmul, fadd, VMLAfd>,
-      Requires<[HasNEON, UseNEONForFP, UseFPVMLx]>;
+      Requires<[HasNEON, UseNEONForFP, UseFPVMLx, NoNEONVFP4]>;
 def : N3VSMulOpPat<fmul, fsub, VMLSfd>,
-      Requires<[HasNEON, UseNEONForFP, UseFPVMLx]>;
+      Requires<[HasNEON, UseNEONForFP, UseFPVMLx, NoNEONVFP4]>;
+def : N3VSMulOpPat<fmul, fadd, VFMAfd>,
+      Requires<[HasNEONVFP4, UseNEONForFP]>;
+def : N3VSMulOpPat<fmul, fsub, VFMSfd>,
+      Requires<[HasNEONVFP4, UseNEONForFP]>;
 def : N2VSPat<fabs, VABSfd>;
 def : N2VSPat<fneg, VNEGfd>;
 def : N3VSPat<NEONfmax, VMAXfd>;