Files
clang-p2996/llvm/test/CodeGen/AMDGPU/trunc-vector-store-assertion-failure.ll
Fangrui Song 9e9907f1cf [AMDGPU,test] Change llc -march= to -mtriple= (#75982)
Similar to 806761a762.

For IR files without a target triple, -mtriple= specifies the full
target triple while -march= merely sets the architecture part of the
default target triple, leaving a target triple which may not make sense,
e.g. amdgpu-apple-darwin.

Therefore, -march= is error-prone and not recommended for tests without
a target triple. The issue has been benign as we recognize
$unknown-apple-darwin as ELF instead of rejecting it outrightly.

This patch changes AMDGPU tests to not rely on the default
OS/environment components. Tests that need fixes are not changed:

```
  LLVM :: CodeGen/AMDGPU/fabs.f64.ll
  LLVM :: CodeGen/AMDGPU/fabs.ll
  LLVM :: CodeGen/AMDGPU/floor.ll
  LLVM :: CodeGen/AMDGPU/fneg-fabs.f64.ll
  LLVM :: CodeGen/AMDGPU/fneg-fabs.ll
  LLVM :: CodeGen/AMDGPU/r600-infinite-loop-bug-while-reorganizing-vector.ll
  LLVM :: CodeGen/AMDGPU/schedule-if-2.ll
```
2024-01-16 21:54:58 -08:00

21 lines
585 B
LLVM

; RUN: llc < %s -mtriple=r600 -mtriple=r600-- -mcpu=redwood | FileCheck %s
; This tests for a bug in the SelectionDAG where custom lowered truncated
; vector stores at the end of a basic block were not being added to the
; LegalizedNodes list, which triggered an assertion failure.
; CHECK-LABEL: {{^}}test:
; CHECK: MEM_RAT_CACHELESS STORE_RAW
define amdgpu_kernel void @test(ptr addrspace(1) %out, i32 %cond, <4 x i8> %in) {
entry:
%0 = icmp eq i32 %cond, 0
br i1 %0, label %if, label %done
if:
store <4 x i8> %in, ptr addrspace(1) %out
br label %done
done:
ret void
}