[ARM] Add patterns for VADD with q and r registers

[oliverlars]

Summary

Added support for VADD to use q and r registers, which reduces pressure on q registers.

[oliverlars]

added run line back and tests for swapping %sp and %src

[dmgreen]

LGTM

[oliverlars]

rG3aed95af4ecd

llvm/lib/Target/ARM/ARMInstrMVE.td

 defm MVE_VADD_qr_i  : MVE_VADDSUB_qr_sizes<"vadd",  "i", 0b0, 0b0, 0b1, 0b0>;
 defm MVE_VQADD_qr_s : MVE_VADDSUB_qr_sizes<"vqadd", "s", 0b1, 0b0, 0b0, 0b0>;
 defm MVE_VQADD_qr_u : MVE_VADDSUB_qr_sizes<"vqadd", "u", 0b1, 0b0, 0b0, 0b1>;
 
 defm MVE_VSUB_qr_i  : MVE_VADDSUB_qr_sizes<"vsub",  "i", 0b0, 0b1, 0b1, 0b0>;
 defm MVE_VQSUB_qr_s : MVE_VADDSUB_qr_sizes<"vqsub", "s", 0b1, 0b1, 0b0, 0b0>;
 defm MVE_VQSUB_qr_u : MVE_VADDSUB_qr_sizes<"vqsub", "u", 0b1, 0b1, 0b0, 0b1>;
 
+let Predicates = [HasMVEInt] in {
+  def : Pat<(v16i8 (add (v16i8 MQPR:$val1), (v16i8 (ARMvdup GPR:$val2)))),
+            (v16i8 (MVE_VADD_qr_i8 (v16i8 MQPR:$val1), (i32 GPR:$val2)))>;
+  def : Pat<(v8i16 (add (v8i16 MQPR:$val1), (v8i16 (ARMvdup GPR:$val2)))),
+            (v8i16 (MVE_VADD_qr_i16 (v8i16 MQPR:$val1), (i32 GPR:$val2)))>;
+  def : Pat<(v4i32 (add (v4i32 MQPR:$val1), (v4i32 (ARMvdup GPR:$val2)))),
+            (v4i32 (MVE_VADD_qr_i32 (v4i32 MQPR:$val1), (i32 GPR:$val2)))>;
+}
+
 class MVE_VQDMULL_qr<string iname, string suffix, bit size,
                      bit T, list<dag> pattern=[]>
   : MVE_qDest_rSrc<iname, suffix, pattern> {
 
   let Inst{28} = size;
   let Inst{21-20} = 0b11;
   let Inst{16} = 0b0;
   let Inst{12} = T;

llvm/test/CodeGen/Thumb2/mve-vaddqr.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+define arm_aapcs_vfpcc <4 x i32> @vaddqr_v4i32(<4 x i32> %src, i32 %src2, <4 x i32> %a, <4 x i32> %b) {
+; CHECK-LABEL: vaddqr_v4i32:
+; CHECK:       @ %bb.0: @ %entry
+; CHECK-NEXT:    vadd.i32 q0, q0, r0
+; CHECK-NEXT:    bx lr
+entry:
+  %i = insertelement <4 x i32> undef, i32 %src2, i32 0
+  %sp = shufflevector <4 x i32> %i, <4 x i32> undef, <4 x i32> zeroinitializer
+  %c = add <4 x i32> %src, %sp
+  ret <4 x i32> %c
+}
+
+define arm_aapcs_vfpcc <8 x i16> @vaddqr_v8i16(<8 x i16> %src, i16 %src2, <8 x i16> %a, <8 x i16> %b) {
+; CHECK-LABEL: vaddqr_v8i16:
+; CHECK:       @ %bb.0: @ %entry
+; CHECK-NEXT:    vadd.i16 q0, q0, r0
+; CHECK-NEXT:    bx lr
+entry:
+  %i = insertelement <8 x i16> undef, i16 %src2, i32 0
+  %sp = shufflevector <8 x i16> %i, <8 x i16> undef, <8 x i32> zeroinitializer
+  %c = add <8 x i16> %src, %sp
+  ret <8 x i16> %c
+}
+
+define arm_aapcs_vfpcc <16 x i8> @vaddqr_v16i8(<16 x i8> %src, i8 %src2, <16 x i8> %a, <16 x i8> %b) {
+; CHECK-LABEL: vaddqr_v16i8:
+; CHECK:       @ %bb.0: @ %entry
+; CHECK-NEXT:    vadd.i8 q0, q0, r0
+; CHECK-NEXT:    bx lr
+entry:
+  %i = insertelement <16 x i8> undef, i8 %src2, i32 0
+  %sp = shufflevector <16 x i8> %i, <16 x i8> undef, <16 x i32> zeroinitializer
+  %c = add <16 x i8> %src, %sp
+  ret <16 x i8> %c
+}
+