AMDGPU: Switch scheduler-subrange-crash.ll to generated checks

Also remove unnecessarily requiring asserts, and replace undef
with poison.

Author: arsenm

llvm/test/CodeGen/AMDGPU/scheduler-subrange-crash.ll

@@ -1,5 +1,5 @@
-; RUN: llc -mtriple=amdgcn < %s | FileCheck %s
-; REQUIRES: asserts
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=amdgcn -mcpu=tonga < %s | FileCheck %s
 ;
 ; This test used to crash with the following assertion:
 ; llc: include/llvm/ADT/IntervalMap.h:632: unsigned int llvm::IntervalMapImpl::LeafNode<llvm::SlotIndex, llvm::LiveInterval *, 8, llvm::IntervalMapInfo<llvm::SlotIndex> >::insertFrom(unsigned int &, unsigned int, KeyT, KeyT, ValT) [KeyT = llvm::SlotIndex, ValT = llvm::LiveInterval *, N = 8, Traits = llvm::IntervalMapInfo<llvm::SlotIndex>]: Assertion `(i == Size || Traits::stopLess(b, start(i))) && "Overlapping insert"' failed.
@@ -8,12 +8,22 @@
 ; (i.e. live interval subranges): subregister defs are not uses for that
 ; purpose.
 ;
-; Check for a valid output:
-; CHECK: tbuffer_store_format_x
-
-target triple = "amdgcn--"
 
 define amdgpu_gs void @main(i32 inreg %arg) #0 {
+; CHECK-LABEL: main:
+; CHECK:       ; %bb.0: ; %main_body
+; CHECK-NEXT:    s_movk_i32 s1, 0x1300
+; CHECK-NEXT:    buffer_load_dword v0, v0, s[0:3], s1 offen
+; CHECK-NEXT:    s_waitcnt vmcnt(0)
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:36 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:48 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:72 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:28 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:64 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:20 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:56 glc slc
+; CHECK-NEXT:    tbuffer_store_format_x v0, off, s[0:3], s0 format:[BUF_DATA_FORMAT_32,BUF_NUM_FORMAT_UINT] offset:92 glc slc
+; CHECK-NEXT:    s_endpgm
 main_body:
   %tmp = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> poison, i32 20, i32 0)
   %tmp1 = call float @llvm.amdgcn.s.buffer.load.f32(<4 x i32> poison, i32 24, i32 0)
@@ -27,17 +37,17 @@ main_body:
   %tmp3 = call i32 @llvm.amdgcn.raw.buffer.load.i32(<4 x i32> poison, i32 poison, i32 4864, i32 0)
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 %tmp3, <4 x i32> poison, i32 36, i32 %arg, i32 68, i32 3)
   %bc = bitcast <4 x float> %array_vector3 to <4 x i32>
-  %tmp4 = extractelement <4 x i32> %bc, i32 undef
+  %tmp4 = extractelement <4 x i32> %bc, i32 poison
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 %tmp4, <4 x i32> poison, i32 48, i32 %arg, i32 68, i32 3)
   %bc49 = bitcast <4 x float> %array_vector11 to <4 x i32>
-  %tmp5 = extractelement <4 x i32> %bc49, i32 undef
+  %tmp5 = extractelement <4 x i32> %bc49, i32 poison
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 %tmp5, <4 x i32> poison, i32 72, i32 %arg, i32 68, i32 3)
   %array_vector21 = insertelement <4 x float> <float 0.000000e+00, float poison, float poison, float poison>, float %tmp, i32 1
   %array_vector22 = insertelement <4 x float> %array_vector21, float poison, i32 2
   %array_vector23 = insertelement <4 x float> %array_vector22, float poison, i32 3
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 poison, <4 x i32> poison, i32 28, i32 %arg, i32 68, i32 3)
   %bc52 = bitcast <4 x float> %array_vector23 to <4 x i32>
-  %tmp6 = extractelement <4 x i32> %bc52, i32 undef
+  %tmp6 = extractelement <4 x i32> %bc52, i32 poison
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 %tmp6, <4 x i32> poison, i32 64, i32 %arg, i32 68, i32 3)
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 poison, <4 x i32> poison, i32 20, i32 %arg, i32 68, i32 3)
   call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 poison, <4 x i32> poison, i32 56, i32 %arg, i32 68, i32 3)
@@ -49,7 +59,7 @@ declare float @llvm.amdgcn.s.buffer.load.f32(<4 x i32>, i32, i32 immarg) #1
 declare i32 @llvm.amdgcn.raw.buffer.load.i32(<4 x i32>, i32, i32, i32 immarg) #2
 declare void @llvm.amdgcn.raw.tbuffer.store.i32(i32, <4 x i32>, i32, i32, i32 immarg, i32 immarg) #3
 
-attributes #0 = { nounwind "target-cpu"="tonga" }
+attributes #0 = { nounwind  }
 attributes #1 = { nounwind readnone willreturn }
 attributes #2 = { nounwind readonly willreturn }
 attributes #3 = { nounwind willreturn writeonly }