Automerge: [RISCV] Add LD_RV32/SD_RV32 to RISCVInstrInfo::canFoldIntoAddrMode. (#140631)

Author: topperc

llvm/lib/Target/RISCV/RISCVInstrInfo.cpp

@@ -2935,6 +2935,7 @@ bool RISCVInstrInfo::canFoldIntoAddrMode(const MachineInstr &MemI, Register Reg,
   case RISCV::LW_INX:
   case RISCV::LWU:
   case RISCV::LD:
+  case RISCV::LD_RV32:
   case RISCV::FLH:
   case RISCV::FLW:
   case RISCV::FLD:
@@ -2944,6 +2945,7 @@ bool RISCVInstrInfo::canFoldIntoAddrMode(const MachineInstr &MemI, Register Reg,
   case RISCV::SW:
   case RISCV::SW_INX:
   case RISCV::SD:
+  case RISCV::SD_RV32:
   case RISCV::FSH:
   case RISCV::FSW:
   case RISCV::FSD:

llvm/test/CodeGen/RISCV/fold-addi-loadstore-zilsd.ll

@@ -28,3 +28,91 @@ entry:
   store double 0.0, ptr @g_0
   ret void
 }
+
+%struct.S = type { double, double }
+
+define double @fold_addi_from_different_bb(i32 %k, i32 %n, ptr %a) nounwind {
+; CHECK-LABEL: fold_addi_from_different_bb:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    addi sp, sp, -32
+; CHECK-NEXT:    sw ra, 28(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    sw s0, 24(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    sw s1, 20(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    sw s2, 16(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    sw s3, 12(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    sw s4, 8(sp) # 4-byte Folded Spill
+; CHECK-NEXT:    blez a1, .LBB2_3
+; CHECK-NEXT:  # %bb.1: # %for.body.lr.ph
+; CHECK-NEXT:    mv s0, a2
+; CHECK-NEXT:    mv s1, a1
+; CHECK-NEXT:    slli a0, a0, 4
+; CHECK-NEXT:    add s4, a2, a0
+; CHECK-NEXT:    fcvt.d.w s2, zero
+; CHECK-NEXT:  .LBB2_2: # %for.body
+; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
+; CHECK-NEXT:    mv a0, s0
+; CHECK-NEXT:    call f
+; CHECK-NEXT:    ld a0, 8(s4)
+; CHECK-NEXT:    addi s1, s1, -1
+; CHECK-NEXT:    fadd.d s2, a0, s2
+; CHECK-NEXT:    bnez s1, .LBB2_2
+; CHECK-NEXT:    j .LBB2_4
+; CHECK-NEXT:  .LBB2_3:
+; CHECK-NEXT:    fcvt.d.w s2, zero
+; CHECK-NEXT:  .LBB2_4: # %for.cond.cleanup
+; CHECK-NEXT:    mv a0, s2
+; CHECK-NEXT:    mv a1, s3
+; CHECK-NEXT:    lw ra, 28(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    lw s0, 24(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    lw s1, 20(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    lw s2, 16(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    lw s3, 12(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    lw s4, 8(sp) # 4-byte Folded Reload
+; CHECK-NEXT:    addi sp, sp, 32
+; CHECK-NEXT:    ret
+entry:
+  %cmp4 = icmp sgt i32 %n, 0
+  br i1 %cmp4, label %for.body.lr.ph, label %for.cond.cleanup
+
+for.body.lr.ph:                                   ; preds = %entry
+  %y = getelementptr inbounds %struct.S, ptr %a, i32 %k, i32 1
+  br label %for.body
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  %s.0.lcssa = phi double [ 0.0, %entry ], [ %add, %for.body ]
+  ret double %s.0.lcssa
+
+for.body:                                         ; preds = %for.body.lr.ph, %for.body
+  %i.06 = phi i32 [ 0, %for.body.lr.ph ], [ %inc, %for.body ]
+  %s.05 = phi double [ 0.0, %for.body.lr.ph ], [ %add, %for.body ]
+  call void @f(ptr %a)
+  %0 = load double, ptr %y, align 8
+  %add = fadd double %0, %s.05
+  %inc = add nuw nsw i32 %i.06, 1
+  %exitcond.not = icmp eq i32 %inc, %n
+  br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
+}
+
+declare void @f(ptr)
+
+define void @split_offset(ptr %dest, double %x) {
+; CHECK-LABEL: split_offset:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    mv a3, a2
+; CHECK-NEXT:    addi a0, a0, 2047
+; CHECK-NEXT:    mv a2, a1
+; CHECK-NEXT:    sd a2, 1(a0)
+; CHECK-NEXT:    sd a2, 9(a0)
+; CHECK-NEXT:    sd a2, 17(a0)
+; CHECK-NEXT:    sd a2, 25(a0)
+; CHECK-NEXT:    ret
+  %p1 = getelementptr double, ptr %dest, i32 256
+  store double %x, ptr %p1
+  %p2 = getelementptr double, ptr %dest, i32 257
+  store double %x, ptr %p2
+  %p3 = getelementptr double, ptr %dest, i32 258
+  store double %x, ptr %p3
+  %p4 = getelementptr double, ptr %dest, i32 259
+  store double %x, ptr %p4
+  ret void
+}