From 7e62684251ed4531cf27e533f0c5df3daf8efb70 Mon Sep 17 00:00:00 2001
From: Simon Pilgrim <llvm-dev@redking.me.uk>
Date: Wed, 8 Apr 2020 13:18:32 +0100
Subject: [PATCH] [AMDGPU] Regenerate vector-extract-insert test checks to fix
 issue reported on D77354

---
 llvm/test/CodeGen/AMDGPU/vector-extract-insert.ll | 106 +++++++++++++++-------
 1 file changed, 75 insertions(+), 31 deletions(-)

diff --git a/llvm/test/CodeGen/AMDGPU/vector-extract-insert.ll b/llvm/test/CodeGen/AMDGPU/vector-extract-insert.ll
index 5bb4159..d87e499 100644
--- a/llvm/test/CodeGen/AMDGPU/vector-extract-insert.ll
+++ b/llvm/test/CodeGen/AMDGPU/vector-extract-insert.ll
@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -march=amdgcn -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 ; Test that when extracting the same unknown vector index from an
@@ -6,14 +7,19 @@
 declare i32 @llvm.amdgcn.workitem.id.x() #0
 
 ; No dynamic indexing required
-; GCN-LABEL: {{^}}extract_insert_same_dynelt_v4i32:
-; GCN: s_load_dword [[VAL:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0xd{{$}}
-; GCN-NOT buffer_load_dword
-; GCN-NOT: [[VAL]]
-; GCN: v_mov_b32_e32 [[VVAL:v[0-9]+]], [[VAL]]
-; GCN-NOT: [[VVAL]]
-; GCN: buffer_store_dword [[VVAL]]
 define amdgpu_kernel void @extract_insert_same_dynelt_v4i32(i32 addrspace(1)* %out, <4 x i32> addrspace(1)* %in, i32 %val, i32 %idx) #1 {
+; GCN-LABEL: extract_insert_same_dynelt_v4i32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT:    s_load_dword s0, s[0:1], 0xd
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_mov_b32 s7, 0xf000
+; GCN-NEXT:    s_mov_b32 s6, 0
+; GCN-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v1, 0
+; GCN-NEXT:    v_mov_b32_e32 v2, s0
+; GCN-NEXT:    buffer_store_dword v2, v[0:1], s[4:7], 0 addr64
+; GCN-NEXT:    s_endpgm
   %id = call i32 @llvm.amdgcn.workitem.id.x()
   %id.ext = sext i32 %id to i64
   %gep.in = getelementptr inbounds <4 x i32>, <4 x i32> addrspace(1)* %in, i64 %id.ext
@@ -25,17 +31,39 @@ define amdgpu_kernel void @extract_insert_same_dynelt_v4i32(i32 addrspace(1)* %o
   ret void
 }
 
-; GCN-LABEL: {{^}}extract_insert_different_dynelt_v4i32:
-; GCN: buffer_load_dwordx4
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: v_cndmask_b32
-; GCN: buffer_store_dword v
 define amdgpu_kernel void @extract_insert_different_dynelt_v4i32(i32 addrspace(1)* %out, <4 x i32> addrspace(1)* %in, i32 %val, i32 %idx0, i32 %idx1) #1 {
+; GCN-LABEL: extract_insert_different_dynelt_v4i32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT:    s_load_dwordx4 s[8:11], s[0:1], 0xd
+; GCN-NEXT:    v_mov_b32_e32 v2, 0
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s2, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_mov_b64 s[0:1], s[6:7]
+; GCN-NEXT:    v_lshlrev_b32_e32 v1, 4, v0
+; GCN-NEXT:    v_lshlrev_b32_e32 v4, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v5, v2
+; GCN-NEXT:    buffer_load_dwordx4 v[0:3], v[1:2], s[0:3], 0 addr64
+; GCN-NEXT:    v_mov_b32_e32 v6, s8
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s9, 3
+; GCN-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_cndmask_b32_e32 v3, v3, v6, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s9, 2
+; GCN-NEXT:    v_cndmask_b32_e32 v2, v2, v6, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s9, 1
+; GCN-NEXT:    v_cndmask_b32_e32 v1, v1, v6, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s9, 0
+; GCN-NEXT:    v_cndmask_b32_e32 v0, v0, v6, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s10, 1
+; GCN-NEXT:    v_cndmask_b32_e32 v0, v0, v1, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s10, 2
+; GCN-NEXT:    v_cndmask_b32_e32 v0, v0, v2, vcc
+; GCN-NEXT:    v_cmp_eq_u32_e64 vcc, s10, 3
+; GCN-NEXT:    v_cndmask_b32_e32 v0, v0, v3, vcc
+; GCN-NEXT:    buffer_store_dword v0, v[4:5], s[4:7], 0 addr64
+; GCN-NEXT:    s_endpgm
   %id = call i32 @llvm.amdgcn.workitem.id.x()
   %id.ext = sext i32 %id to i64
   %gep.in = getelementptr inbounds <4 x i32>, <4 x i32> addrspace(1)* %in, i64 %id.ext
@@ -47,14 +75,19 @@ define amdgpu_kernel void @extract_insert_different_dynelt_v4i32(i32 addrspace(1
   ret void
 }
 
-; GCN-LABEL: {{^}}extract_insert_same_elt2_v4i32:
-; GCN: s_load_dword [[VAL:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0xd{{$}}
-; GCN-NOT buffer_load_dword
-; GCN-NOT: [[VAL]]
-; GCN: v_mov_b32_e32 [[VVAL:v[0-9]+]], [[VAL]]
-; GCN-NOT: [[VVAL]]
-; GCN: buffer_store_dword [[VVAL]]
 define amdgpu_kernel void @extract_insert_same_elt2_v4i32(i32 addrspace(1)* %out, <4 x i32> addrspace(1)* %in, i32 %val, i32 %idx) #1 {
+; GCN-LABEL: extract_insert_same_elt2_v4i32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT:    s_load_dword s0, s[0:1], 0xd
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_mov_b32 s7, 0xf000
+; GCN-NEXT:    s_mov_b32 s6, 0
+; GCN-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v1, 0
+; GCN-NEXT:    v_mov_b32_e32 v2, s0
+; GCN-NEXT:    buffer_store_dword v2, v[0:1], s[4:7], 0 addr64
+; GCN-NEXT:    s_endpgm
   %id = call i32 @llvm.amdgcn.workitem.id.x()
   %id.ext = sext i32 %id to i64
   %gep.in = getelementptr inbounds <4 x i32>, <4 x i32> addrspace(1)* %in, i64 %id.ext
@@ -66,14 +99,25 @@ define amdgpu_kernel void @extract_insert_same_elt2_v4i32(i32 addrspace(1)* %out
   ret void
 }
 
-; GCN-LABEL: {{^}}extract_insert_same_dynelt_v4f32:
-; GCN: s_load_dword [[VAL:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0xd{{$}}
-; GCN-NOT buffer_load_dword
-; GCN-NOT: [[VAL]]
-; GCN: v_mov_b32_e32 [[VVAL:v[0-9]+]], [[VAL]]
-; GCN-NOT: [[VVAL]]
-; GCN: buffer_store_dword [[VVAL]]
 define amdgpu_kernel void @extract_insert_same_dynelt_v4f32(float addrspace(1)* %out, <4 x float> addrspace(1)* %in, float %val, i32 %idx) #1 {
+; GCN-LABEL: extract_insert_same_dynelt_v4f32:
+; GCN:       ; %bb.0:
+; GCN-NEXT:    s_load_dwordx4 s[4:7], s[0:1], 0x9
+; GCN-NEXT:    s_load_dword s8, s[0:1], 0xd
+; GCN-NEXT:    v_mov_b32_e32 v2, 0
+; GCN-NEXT:    s_mov_b32 s3, 0xf000
+; GCN-NEXT:    s_mov_b32 s2, 0
+; GCN-NEXT:    s_waitcnt lgkmcnt(0)
+; GCN-NEXT:    s_mov_b64 s[0:1], s[6:7]
+; GCN-NEXT:    v_lshlrev_b32_e32 v1, 4, v0
+; GCN-NEXT:    v_lshlrev_b32_e32 v4, 2, v0
+; GCN-NEXT:    v_mov_b32_e32 v5, v2
+; GCN-NEXT:    buffer_load_dwordx4 v[0:3], v[1:2], s[0:3], 0 addr64
+; GCN-NEXT:    s_mov_b64 s[6:7], s[2:3]
+; GCN-NEXT:    s_waitcnt vmcnt(0)
+; GCN-NEXT:    v_mov_b32_e32 v0, s8
+; GCN-NEXT:    buffer_store_dword v0, v[4:5], s[4:7], 0 addr64
+; GCN-NEXT:    s_endpgm
   %id = call i32 @llvm.amdgcn.workitem.id.x()
   %id.ext = sext i32 %id to i64
   %gep.in = getelementptr inbounds <4 x float>, <4 x float> addrspace(1)* %in, i64 %id.ext
-- 
2.7.4