; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc < %s -mtriple=riscv64 -mattr=+v -verify-machineinstrs | FileCheck %s

; Check that EVL tail folded loops from the loop vectorizer are able to have the
; vl of non-VP instructions reduced.
define void @evl_tail_folded(ptr %p, ptr %q) {
; CHECK-LABEL: evl_tail_folded:
; CHECK:       # %bb.0: # %entry
; CHECK-NEXT:    li a2, 0
; CHECK-NEXT:    csrr a3, vlenb
; CHECK-NEXT:    srli a3, a3, 2
; CHECK-NEXT:    addi a4, a3, 1023
; CHECK-NEXT:    neg a5, a3
; CHECK-NEXT:    and a4, a4, a5
; CHECK-NEXT:    li a5, 1024
; CHECK-NEXT:  .LBB0_1: # %vector.body
; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    sub a6, a5, a2
; CHECK-NEXT:    slli a7, a2, 3
; CHECK-NEXT:    vsetvli a6, a6, e64, m2, ta, ma
; CHECK-NEXT:    add t0, a0, a7
; CHECK-NEXT:    vle64.v v8, (t0)
; CHECK-NEXT:    sub a4, a4, a3
; CHECK-NEXT:    add a7, a1, a7
; CHECK-NEXT:    vadd.vi v8, v8, 1
; CHECK-NEXT:    vse64.v v8, (a7)
; CHECK-NEXT:    add a2, a2, a6
; CHECK-NEXT:    bnez a4, .LBB0_1
; CHECK-NEXT:  # %bb.2: # %exit
; CHECK-NEXT:    ret
entry:
  %0 = tail call i64 @llvm.vscale.i64()
  %1 = shl i64 %0, 1
  %n.rnd.up = add i64 %1, 1023
  %n.mod.vf = urem i64 %n.rnd.up, %1
  %n.vec = sub i64 %n.rnd.up, %n.mod.vf
  %2 = tail call i64 @llvm.vscale.i64()
  %3 = shl i64 %2, 1
  br label %vector.body

vector.body:
  %index = phi i64 [ 0, %entry ], [ %index.next, %vector.body ]
  %evl.based.iv = phi i64 [ 0, %entry ], [ %index.evl.next, %vector.body ]
  %avl = sub i64 1024, %evl.based.iv
  %4 = tail call i32 @llvm.experimental.get.vector.length.i64(i64 %avl, i32 2, i1 true)
  %5 = getelementptr i64, ptr %p, i64 %evl.based.iv
  %vp.op.load = tail call <vscale x 2 x i64> @llvm.vp.load.nxv2i64.p0(ptr %5, <vscale x 2 x i1> splat (i1 true), i32 %4)
  %6 = add <vscale x 2 x i64> %vp.op.load, splat (i64 1)
  %7 = getelementptr i64, ptr %q, i64 %evl.based.iv
  tail call void @llvm.vp.store.nxv2i64.p0(<vscale x 2 x i64> %6, ptr %7, <vscale x 2 x i1> splat (i1 true), i32 %4)
  %8 = zext i32 %4 to i64
  %index.evl.next = add i64 %evl.based.iv, %8
  %index.next = add i64 %index, %3
  %9 = icmp eq i64 %index.next, %n.vec
  br i1 %9, label %exit, label %vector.body

exit:
  ret void
}