derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
110 lines
3.1 KiB
ArmAsm
110 lines
3.1 KiB
ArmAsm
# This test checks that branch inversion works when reordering blocks which
|
|
# contain short range conditional branches. Handles edge cases, like when
|
|
# the immediate value is the upper or lower allowed value in which case the
|
|
# transformation bails.
|
|
|
|
# REQUIRES: system-linux, asserts
|
|
|
|
# RUN: %clang %cflags -march=armv9-a+cmpbr -Wl,-q %s -o %t
|
|
# RUN: link_fdata --no-lbr %s %t %t.fdata
|
|
# RUN: llvm-strip --strip-unneeded %t
|
|
# RUN: llvm-bolt -v=1 %t -o %t.bolt --data %t.fdata --reorder-blocks=ext-tsp --compact-code-model \
|
|
# RUN: | FileCheck %s --check-prefix=BOLT-INFO
|
|
# RUN: llvm-objdump -d %t.bolt | FileCheck %s
|
|
|
|
# CHECK: Disassembly of section .text:
|
|
|
|
.globl immediate_increment
|
|
.type immediate_increment, %function
|
|
immediate_increment:
|
|
.entry0:
|
|
# FDATA: 1 immediate_increment #.entry0# 10
|
|
cbgt x0, #0, .exit0
|
|
.cold0:
|
|
# FDATA: 1 immediate_increment #.cold0# 1
|
|
mov x0, #1
|
|
ret
|
|
.exit0:
|
|
# FDATA: 1 immediate_increment #.exit0# 10
|
|
mov x0, #2
|
|
ret
|
|
|
|
# CHECK: <immediate_increment>:
|
|
# CHECK-NEXT: {{.*}} cblt x0, #0x1, 0x[[ADDR0:[0-9a-f]+]] <{{.*}}>
|
|
# CHECK-NEXT: {{.*}} mov x0, #0x2 // =2
|
|
# CHECK-NEXT: {{.*}} ret
|
|
# CHECK-NEXT: [[ADDR0]]: {{.*}} mov x0, #0x1 // =1
|
|
# CHECK-NEXT: {{.*}} ret
|
|
|
|
.globl immediate_decrement
|
|
.type immediate_decrement, %function
|
|
immediate_decrement:
|
|
.entry1:
|
|
# FDATA: 1 immediate_decrement #.entry1# 10
|
|
cblo x0, #1, .exit1
|
|
.cold1:
|
|
# FDATA: 1 immediate_decrement #.cold1# 1
|
|
mov x0, #1
|
|
ret
|
|
.exit1:
|
|
# FDATA: 1 immediate_decrement #.exit1# 10
|
|
mov x0, #2
|
|
ret
|
|
|
|
# CHECK: <immediate_decrement>:
|
|
# CHECK-NEXT: {{.*}} cbhi x0, #0x0, 0x[[ADDR1:[0-9a-f]+]] <{{.*}}>
|
|
# CHECK-NEXT: {{.*}} mov x0, #0x2 // =2
|
|
# CHECK-NEXT: {{.*}} ret
|
|
# CHECK-NEXT: [[ADDR1]]: {{.*}} mov x0, #0x1 // =1
|
|
# CHECK-NEXT: {{.*}} ret
|
|
|
|
.globl register_swap
|
|
.type register_swap, %function
|
|
register_swap:
|
|
.entry2:
|
|
# FDATA: 1 register_swap #.entry2# 10
|
|
cbge x0, x1, .exit2
|
|
.cold2:
|
|
# FDATA: 1 register_swap #.cold2# 1
|
|
mov x0, #1
|
|
ret
|
|
.exit2:
|
|
# FDATA: 1 register_swap #.exit2# 10
|
|
mov x0, #2
|
|
ret
|
|
|
|
# CHECK: <register_swap>:
|
|
# CHECK-NEXT: {{.*}} cbgt x1, x0, 0x[[ADDR2:[0-9a-f]+]] <{{.*}}>
|
|
# CHECK-NEXT: {{.*}} mov x0, #0x2 // =2
|
|
# CHECK-NEXT: {{.*}} ret
|
|
# CHECK-NEXT: [[ADDR2]]: {{.*}} mov x0, #0x1 // =1
|
|
# CHECK-NEXT: {{.*}} ret
|
|
|
|
.globl irreversible
|
|
.type irreversible, %function
|
|
irreversible:
|
|
.entry3:
|
|
# FDATA: 1 irreversible #.entry3# 10
|
|
cbgt x0, #63, .exit3
|
|
.cold3:
|
|
# FDATA: 1 irreversible #.cold3# 1
|
|
mov x0, #1
|
|
ret
|
|
.exit3:
|
|
# FDATA: 1 irreversible #.exit3# 10
|
|
mov x0, #2
|
|
ret
|
|
|
|
# BOLT-INFO: unable to swap successors in irreversible
|
|
|
|
# CHECK: <irreversible>:
|
|
# CHECK-NEXT: {{.*}} cbgt x0, #0x3f, 0x[[ADDR3:[0-9a-f]+]] <{{.*}}>
|
|
# CHECK-NEXT: {{.*}} b 0x[[ADDR4:[0-9a-f]+]] <{{.*}}>
|
|
# CHECK-NEXT: [[ADDR3]]: {{.*}} mov x0, #0x2 // =2
|
|
# CHECK-NEXT: {{.*}} ret
|
|
# CHECK-NEXT: [[ADDR4]]: {{.*}} mov x0, #0x1 // =1
|
|
# CHECK-NEXT: {{.*}} ret
|
|
|
|
## Force relocation mode.
|
|
.reloc 0, R_AARCH64_NONE
|