Currently assembler accepts null for 32-bit and shorter destination operands. Null should be enabled for 64-bit dst operands as well.
Examples of failed tests:
s_or_b64 null, s[0:1], s[0:1] s_atomic_add_x2 null, s[0:1], s0
Paths
| Differential D68785
[AMDGPU][MC][GFX10] Enabled null for 64-bit dst operands ClosedPublic Authored by dp on Oct 10 2019, 5:18 AM.
Details Summary Currently assembler accepts null for 32-bit and shorter destination operands. Null should be enabled for 64-bit dst operands as well. Examples of failed tests: s_or_b64 null, s[0:1], s[0:1] s_atomic_add_x2 null, s[0:1], s0
Diff Detail
Event TimelineHerald added subscribers: jfb, t-tye, tpr and 6 others. · View Herald TranscriptOct 10 2019, 5:18 AM This revision is now accepted and ready to land.Oct 10 2019, 8:00 AM Closed by commit rG472c6b0aa022: [AMDGPU][MC][GFX10] Enabled null for 64-bit dst operands (authored by dp). · Explain WhyOct 11 2019, 7:36 AM This revision was automatically updated to reflect the committed changes.
Revision Contents
Diff 224592 llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
llvm/test/MC/AMDGPU/sop1.s
llvm/test/MC/AMDGPU/sop2.s
llvm/test/MC/AMDGPU/sopk.s
llvm/test/MC/Disassembler/AMDGPU/gfx10_dasm_all.txt
|