Mercurial > hg > CbC > CbC_llvm
comparison llvm/test/CodeGen/AMDGPU/at-least-one-def-value-assert.mir @ 150:1d019706d866
LLVM10
author | anatofuz |
---|---|
date | Thu, 13 Feb 2020 15:10:13 +0900 (2020-02-13) |
parents | |
children | 0572611fdcc8 |
comparison
equal
deleted
inserted
replaced
147:c2174574ed3a | 150:1d019706d866 |
---|---|
1 # RUN: not llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx906 -verify-machineinstrs -run-pass=machine-scheduler -verify-misched -o /dev/null %s 2>&1 | FileCheck %s | |
2 | |
3 # CHECK: *** Bad machine code: No live subrange at use *** | |
4 # CHECK-NEXT: - function: at_least_one_value_should_be_defined_by_this_mask | |
5 # CHECK-NEXT: - basic block: %bb.0 | |
6 # CHECK-NEXT: - instruction: 48B dead undef %2.sub0:vreg_128 = COPY %0.sub0:vreg_128 | |
7 # CHECK-NEXT: - operand 1: %0.sub0:vreg_128 | |
8 # CHECK-NEXT: - interval: %0 [16r,48r:0) 0@16r L00000002 [16r,32r:0) 0@16r weight:0.000000e+00 | |
9 | |
10 # This used to assert with: !SR.empty() && "At least one value should be defined by this mask" | |
11 | |
12 # This MIR is invalid and should be caught by the verifier. %0.sub0 is | |
13 # used, but not defined. There are also lanes in %0 that are not used | |
14 # or defined anywhere. Previously there was an assertion in the | |
15 # LiveInterval computation, which was more confusing. The invalid | |
16 # LiveRange should be produced and the verifier will catch it. | |
17 | |
18 --- | |
19 name: at_least_one_value_should_be_defined_by_this_mask | |
20 tracksRegLiveness: true | |
21 body: | | |
22 bb.0: | |
23 | |
24 undef %0.sub1:vreg_128 = V_MOV_B32_e32 0, implicit $exec | |
25 %1:vreg_128 = COPY %0 | |
26 undef %2.sub0:vreg_128 = COPY %0.sub0 | |
27 | |
28 ... |