VLENB is the length of a vector register in bytes. We use
<vscale x 64 bits> to represent one vector register. The dwarf offset is
VLENB * scalable_offset / 8.
For the mask vector, it occupies one vector register.
Paths
| Differential D107432
[RISCV] Emit DWARF location expression for RVV stack objects. ClosedPublic Authored by HsiangKai on Aug 4 2021, 1:19 AM.
Details Summary VLENB is the length of a vector register in bytes. We use For the mask vector, it occupies one vector register.
Diff Detail
Event TimelineHerald added subscribers: vkmr, frasercrmck, evandro and 25 others. · View Herald TranscriptAug 4 2021, 1:19 AM Comment Actions Sorry, I'm a bit out of my depth with these DWARF offset opcodes. I think what you're doing here makes sense, though. Herald added subscribers: VincentWu, achieveartificialintelligence. · View Herald TranscriptNov 5 2021, 6:24 PM
This revision is now accepted and ready to land.Nov 23 2021, 4:18 AM Closed by commit rGb0c742152489: [RISCV] Emit DWARF location expression for RVV stack objects. (authored by HsiangKai). · Explain WhyNov 26 2021, 11:13 PM This revision was automatically updated to reflect the committed changes.
Revision Contents
Diff 363996 llvm/lib/Target/RISCV/RISCVRegisterInfo.h
llvm/lib/Target/RISCV/RISCVRegisterInfo.cpp
llvm/lib/Target/RISCV/RISCVRegisterInfo.td
llvm/lib/Target/RISCV/RISCVSystemOperands.td
llvm/test/CodeGen/RISCV/rvv/debug-info-rvv-dbg-value.mir
|
Is this path tested by this patch? I don't see any DW_OP_minus in the expected output. I may be missing something, though.