[llvm] [LV][EVL]Support reversed loads/stores. (PR #88025)

[via llvm-commits]

================
@@ -117,3 +113,147 @@ for.body:
 loopend:
   ret void
 }
+
+define void @reverse_load_store_masked(i64 %startval, ptr noalias %ptr, ptr noalias %ptr1, ptr noalias %ptr2) {
+; IF-EVL-LABEL: @reverse_load_store_masked(
+; IF-EVL-NEXT:  entry:
+; IF-EVL-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; IF-EVL:       vector.ph:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = mul i64 [[TMP0]], 4
+; IF-EVL-NEXT:    [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP3:%.*]] = mul i64 [[TMP2]], 4
+; IF-EVL-NEXT:    [[TMP4:%.*]] = sub i64 [[TMP3]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 1024, [[TMP4]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP1]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[IND_END:%.*]] = sub i64 [[STARTVAL:%.*]], [[N_VEC]]
+; IF-EVL-NEXT:    [[IND_END1:%.*]] = trunc i64 [[N_VEC]] to i32
+; IF-EVL-NEXT:    [[TMP5:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP6:%.*]] = mul i64 [[TMP5]], 4
+; IF-EVL-NEXT:    br label [[VECTOR_BODY:%.*]]
+; IF-EVL:       vector.body:
+; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], [[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP7:%.*]] = sub i64 1024, [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP8:%.*]] = call i32 @llvm.experimental.get.vector.length.i64(i64 [[TMP7]], i32 4, i1 true)
+; IF-EVL-NEXT:    [[OFFSET_IDX:%.*]] = sub i64 [[STARTVAL]], [[EVL_BASED_IV]]
+; IF-EVL-NEXT:    [[TMP9:%.*]] = add i64 [[OFFSET_IDX]], 0
+; IF-EVL-NEXT:    [[OFFSET_IDX3:%.*]] = trunc i64 [[EVL_BASED_IV]] to i32
+; IF-EVL-NEXT:    [[TMP10:%.*]] = add i32 [[OFFSET_IDX3]], 0
+; IF-EVL-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x i64> poison, i64 [[EVL_BASED_IV]], i64 0
+; IF-EVL-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x i64> [[BROADCAST_SPLATINSERT]], <vscale x 4 x i64> poison, <vscale x 4 x i32> zeroinitializer
+; IF-EVL-NEXT:    [[TMP11:%.*]] = call <vscale x 4 x i64> @llvm.experimental.stepvector.nxv4i64()
+; IF-EVL-NEXT:    [[TMP12:%.*]] = add <vscale x 4 x i64> zeroinitializer, [[TMP11]]
+; IF-EVL-NEXT:    [[VEC_IV:%.*]] = add <vscale x 4 x i64> [[BROADCAST_SPLAT]], [[TMP12]]
+; IF-EVL-NEXT:    [[TMP13:%.*]] = icmp ule <vscale x 4 x i64> [[VEC_IV]], shufflevector (<vscale x 4 x i64> insertelement (<vscale x 4 x i64> poison, i64 1023, i64 0), <vscale x 4 x i64> poison, <vscale x 4 x i32> zeroinitializer)
+; IF-EVL-NEXT:    [[TMP14:%.*]] = add i64 [[TMP9]], -1
+; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, ptr [[PTR:%.*]], i32 [[TMP10]]
+; IF-EVL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i32, ptr [[TMP15]], i32 0
+; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.vp.load.nxv4i32.p0(ptr align 4 [[TMP16]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP8]])
+; IF-EVL-NEXT:    [[TMP17:%.*]] = icmp slt <vscale x 4 x i32> [[VP_OP_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 100, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
+; IF-EVL-NEXT:    [[TMP18:%.*]] = select <vscale x 4 x i1> [[TMP13]], <vscale x 4 x i1> [[TMP17]], <vscale x 4 x i1> zeroinitializer
+; IF-EVL-NEXT:    [[TMP19:%.*]] = getelementptr i32, ptr [[PTR1:%.*]], i64 [[TMP14]]
+; IF-EVL-NEXT:    [[TMP20:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP21:%.*]] = mul i64 [[TMP20]], 4
+; IF-EVL-NEXT:    [[TMP22:%.*]] = mul i64 0, [[TMP21]]
+; IF-EVL-NEXT:    [[TMP23:%.*]] = sub i64 1, [[TMP21]]
+; IF-EVL-NEXT:    [[TMP24:%.*]] = getelementptr i32, ptr [[TMP19]], i64 [[TMP22]]
+; IF-EVL-NEXT:    [[TMP25:%.*]] = getelementptr i32, ptr [[TMP24]], i64 [[TMP23]]
+; IF-EVL-NEXT:    [[VP_REVERSE_MASK:%.*]] = call <vscale x 4 x i1> @llvm.experimental.vp.reverse.nxv4i1(<vscale x 4 x i1> [[TMP18]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP8]])
+; IF-EVL-NEXT:    [[VP_OP_LOAD4:%.*]] = call <vscale x 4 x i32> @llvm.vp.load.nxv4i32.p0(ptr align 4 [[TMP25]], <vscale x 4 x i1> [[VP_REVERSE_MASK]], i32 [[TMP8]])
+; IF-EVL-NEXT:    [[VP_REVERSE:%.*]] = call <vscale x 4 x i32> @llvm.experimental.vp.reverse.nxv4i32(<vscale x 4 x i32> [[VP_OP_LOAD4]], <vscale x 4 x i1> shufflevector (<vscale x 4 x i1> insertelement (<vscale x 4 x i1> poison, i1 true, i64 0), <vscale x 4 x i1> poison, <vscale x 4 x i32> zeroinitializer), i32 [[TMP8]])
----------------
ayalz wrote:

Should the above reverse of the loaded values use the partial mask of these values, rather than an all-true mask?

https://github.com/llvm/llvm-project/pull/88025