1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 19:52:54 +01:00
llvm-mirror/test/CodeGen/Hexagon/newvaluejump-float.mir
Krzysztof Parzyszek 660a77777d [Hexagon] Don't form new-value jumps from floating-point instructions
Additionally, verify that the register defined by the producer is a
32-bit register.

llvm-svn: 324381
2018-02-06 19:08:41 +00:00

20 lines
437 B
YAML

# RUN: llc -march=hexagon -run-pass=hexagon-nvj %s -o - | FileCheck %s
# Check that we don't generate a new-value jump for a floating-point
# instruction.
# CHECK-NOT: J4_cmpgti_t_jumpnv_t
---
name: fred
tracksRegLiveness: true
body: |
bb.0:
liveins: $d0
$r0 = F2_conv_df2w_chop $d0, implicit $usr
$p0 = C2_cmpgti $r0, 30
J2_jumpt $p0, %bb.1, implicit-def $pc
bb.1:
J2_jumpr $r31, implicit-def $pc
...