This is an archive of the discontinued LLVM Phabricator instance.

Differential D146934

[AMDGPU] Update mcp-overlap-after-propagation.mir test
ClosedPublic

Authored by barannikov88 on Mar 26 2023, 11:27 PM.

Details

Reviewers
arsenm
tpr
Commits
rG790252dfb372: [AMDGPU] Update mcp-overlap-after-propagation.mir test
Summary

The test was added in D69953, but since D67794 landed a bit later it no
longer catches the bug. Update the test to be representative again.

Diff Detail

Event Timeline

barannikov88 created this revision.Mar 26 2023, 11:27 PM
Herald added a project: Restricted Project. · View Herald TranscriptMar 26 2023, 11:27 PM
Herald added subscribers: kosarev, kerbowa, tpr and 4 others. · View Herald Transcript
barannikov88 requested review of this revision.Mar 26 2023, 11:27 PM
Herald added a project: Restricted Project. · View Herald TranscriptMar 26 2023, 11:27 PM
Herald added subscribers: llvm-commits, wdng. · View Herald Transcript
barannikov88 added reviewers: arsenm, tpr.Mar 26 2023, 11:28 PM
barannikov88 added a comment.Mar 26 2023, 11:33 PM

The change effectively disables BackwardCopyPropagateBlock that folded the first copy into
the preceding load, eliminating definition of s[60-67], which is necessary to reproduce the bug.

Harbormaster completed remote builds in B221890: Diff 508491.Mar 27 2023, 12:33 AM
arsenm accepted this revision.Apr 4 2023, 7:18 AM
This revision is now accepted and ready to land.Apr 4 2023, 7:18 AM
Closed by commit rG790252dfb372: [AMDGPU] Update mcp-overlap-after-propagation.mir test (authored by barannikov88). · Explain WhyApr 8 2023, 3:10 PM
This revision was automatically updated to reflect the committed changes.
barannikov88 added a commit: rG790252dfb372: [AMDGPU] Update mcp-overlap-after-propagation.mir test.

Revision Contents

PathSize
llvm/
test/
CodeGen/
AMDGPU/
6 lines

Diff 511930

llvm/test/CodeGen/AMDGPU/mcp-overlap-after-propagation.mir

# RUN: llc -march=amdgcn -mcpu=gfx1010 %s -o - -run-pass machine-cp -verify-machineinstrs | FileCheck %s # RUN: llc -march=amdgcn -mcpu=gfx1010 %s -o - -run-pass machine-cp -verify-machineinstrs | FileCheck %s
# RUN: llc -march=amdgcn -mcpu=gfx1100 %s -o - -run-pass machine-cp -verify-machineinstrs | FileCheck %s # RUN: llc -march=amdgcn -mcpu=gfx1100 %s -o - -run-pass machine-cp -verify-machineinstrs | FileCheck %s
# #
# The MachineCopyPropagation bug being tested propagates s[60:67] into the copy # The MachineCopyPropagation bug being tested propagates s[60:67] into the copy
# into s[56:63], and then uses s[60:67] in the following # into s[56:63], and then uses s[60:67] in the following
# IMAGE_SAMPLE_V3_V2_gfx10, even though it has just overwritten half of it. # IMAGE_SAMPLE_V3_V2_gfx10, even though it has just overwritten half of it.
# CHECK-LABEL: name: _amdgpu_ps_main # CHECK-LABEL: name: _amdgpu_ps_main
# CHECK-NOT: IMAGE_SAMPLE_V3_V2_gfx10 {{.*}} $sgpr60_sgpr61_sgpr62_sgpr63_sgpr64_sgpr65_sgpr66_sgpr67 # CHECK-NOT: IMAGE_SAMPLE_V3_V2_gfx10 {{.*}} $sgpr60_sgpr61_sgpr62_sgpr63_sgpr64_sgpr65_sgpr66_sgpr67
# CHECK: IMAGE_SAMPLE_V3_V2_gfx10 {{.*}} $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95 # CHECK: IMAGE_SAMPLE_V3_V2_gfx10 {{.*}} $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
--- ---
name: _amdgpu_ps_main name: _amdgpu_ps_main
body: | body: |
bb.0: bb.0:
successors: successors:
liveins: $sgpr2, $sgpr3, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr5, $vgpr70, $vgpr71 liveins: $sgpr60, $sgpr61, $sgpr62, $sgpr63, $sgpr64, $sgpr65, $sgpr66, $sgpr67, $sgpr96, $sgpr97, $sgpr98, $sgpr99, $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr5, $vgpr70, $vgpr71
renamable $sgpr8_sgpr9 = S_GETPC_B64
renamable $sgpr8 = COPY killed renamable $sgpr2
renamable $sgpr60_sgpr61_sgpr62_sgpr63_sgpr64_sgpr65_sgpr66_sgpr67 = S_LOAD_DWORDX8_IMM renamable $sgpr8_sgpr9, 144, 0 :: (invariant load (s256), align 16, addrspace 4)
renamable $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95 = COPY killed renamable $sgpr60_sgpr61_sgpr62_sgpr63_sgpr64_sgpr65_sgpr66_sgpr67 renamable $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95 = COPY killed renamable $sgpr60_sgpr61_sgpr62_sgpr63_sgpr64_sgpr65_sgpr66_sgpr67
renamable $vgpr4 = IMAGE_GET_LOD_V1_V2_gfx10 renamable $vgpr70_vgpr71, renamable $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95, renamable $sgpr96_sgpr97_sgpr98_sgpr99, 2, 1, 0, 0, 0, 0, 0, 0, implicit $exec
renamable $sgpr56_sgpr57_sgpr58_sgpr59_sgpr60_sgpr61_sgpr62_sgpr63 = COPY killed renamable $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95 renamable $sgpr56_sgpr57_sgpr58_sgpr59_sgpr60_sgpr61_sgpr62_sgpr63 = COPY killed renamable $sgpr88_sgpr89_sgpr90_sgpr91_sgpr92_sgpr93_sgpr94_sgpr95
renamable $vgpr12_vgpr13_vgpr14 = IMAGE_SAMPLE_V3_V2_gfx10 renamable $vgpr70_vgpr71, renamable $sgpr56_sgpr57_sgpr58_sgpr59_sgpr60_sgpr61_sgpr62_sgpr63, renamable $sgpr96_sgpr97_sgpr98_sgpr99, 7, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s96), align 16) renamable $vgpr12_vgpr13_vgpr14 = IMAGE_SAMPLE_V3_V2_gfx10 renamable $vgpr70_vgpr71, renamable $sgpr56_sgpr57_sgpr58_sgpr59_sgpr60_sgpr61_sgpr62_sgpr63, renamable $sgpr96_sgpr97_sgpr98_sgpr99, 7, 1, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s96), align 16)
S_ENDPGM 0 S_ENDPGM 0
... ...