Similar to 806761a762
-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.
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.
25 lines
733 B
YAML
25 lines
733 B
YAML
# RUN: llc -mtriple=hexagon -mcpu=hexagonv62 -run-pass amode-opt %s -o -
|
|
# REQUIRES: asserts
|
|
#
|
|
# This test merely checks if the pass that optimizes addressing modes in the
|
|
# hexagon backend doesn't crash when the following code is presented to it.
|
|
#
|
|
# This is a testcase reduced from code generated by the Halide compiler for a
|
|
# Halide pipeline. This code was causing the pass 'amode-opt' to crash because
|
|
# it was looking for the third operand (offset) in the following instruction
|
|
# that does not, in fact, have a third operand.
|
|
#
|
|
# $r1 = L2_loadw_locked $r1
|
|
#
|
|
# Check that this doesn't crash.
|
|
|
|
|
|
name: fred
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $r0
|
|
$r1 = A2_addi $r0, 4
|
|
$r1 = L2_loadw_locked $r1
|
|
...
|