| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | [CodeGen] Treat reserved registers as always live | Richard-Qin-X | 2026-08-11 | 2 | -0/+52 |
| | | |||||
| * | [CodeGen] Accept conservative physical live-ins | Richard-Qin-X | 2026-08-11 | 2 | -2/+8 |
| | | |||||
| * | [RISCV] Model typed floating-point register views | Richard-Qin-X | 2026-08-11 | 14 | -155/+303 |
| | | |||||
| * | [RISCV] Order call operands like LLVM MIR | Richard-Qin-X | 2026-08-11 | 2 | -15/+28 |
| | | |||||
| * | [RISCV] Add LP64 calling-convention facts | Richard-Qin-X | 2026-08-11 | 9 | -69/+146 |
| | | |||||
| * | [CodeGen] Remove redundant numeric call-site IDs | Richard-Qin-X | 2026-08-11 | 10 | -46/+18 |
| | | |||||
| * | [RISCV] Derive the generated instruction count in tests | Richard-Qin-X | 2026-08-11 | 1 | -1/+5 |
| | | |||||
| * | [RISCV] Match LLVM call-frame pseudo semantics | Richard-Qin-X | 2026-08-11 | 6 | -85/+80 |
| | | |||||
| * | [RISCV] Use canonical copies for full ABI values | Richard-Qin-X | 2026-08-11 | 2 | -7/+15 |
| | | |||||
| * | [RISCV] Lower multiply without M through libcalls | Richard-Qin-X | 2026-08-11 | 2 | -3/+182 |
| | | |||||
| * | [RISCV] Accept generic scalars in call lowering | Richard-Qin-X | 2026-08-11 | 2 | -3/+18 |
| | | |||||
| * | [RISCV] Honor invocation subtarget features | Richard-Qin-X | 2026-08-11 | 10 | -18/+230 |
| | | |||||
| * | [RISCV] Fold narrow binary artifact chains | Richard-Qin-X | 2026-08-11 | 2 | -15/+118 |
| | | |||||
| * | [CodeGen] Revisit instructions invalidated by legalization | Richard-Qin-X | 2026-08-11 | 2 | -12/+72 |
| | | |||||
| * | [RISCV] Preserve folded constant placement | Richard-Qin-X | 2026-08-11 | 2 | -3/+38 |
| | | |||||
| * | [RISCV] Reuse compare operand extensions | Richard-Qin-X | 2026-08-11 | 2 | -0/+71 |
| | | |||||
| * | [RISCV] Legalize integer compares to RV64 containers | Richard-Qin-X | 2026-08-11 | 2 | -20/+204 |
| | | |||||
| * | [RISCV] Preserve live constants when folding extensions | Richard-Qin-X | 2026-08-11 | 2 | -2/+52 |
| | | |||||
| * | [MIR] Omit absent zero-offset MMO bases | Richard-Qin-X | 2026-08-11 | 2 | -5/+14 |
| | | |||||
| * | [RISCV] Lower cmpxchg with success | Richard-Qin-X | 2026-08-11 | 2 | -1/+131 |
| | | |||||
| * | [RISCV] Preserve narrow cmpxchg results when widening | Richard-Qin-X | 2026-08-11 | 2 | -2/+40 |
| | | |||||
| * | [RISCV] Legalize narrow atomic compare exchange | Richard-Qin-X | 2026-08-11 | 3 | -2/+228 |
| | | |||||
| * | [MIR] Parse one-ordering compare-exchange MMOs | Richard-Qin-X | 2026-08-11 | 4 | -52/+94 |
| | | |||||
| * | [RISCV] Legalize fixed-baseline RV64 MachineIR | Richard-Qin-X | 2026-08-11 | 18 | -65/+3013 |
| | | |||||
| * | [MIR] Accept fixed-baseline LLVM syntax | Richard-Qin-X | 2026-08-11 | 53 | -650/+2429 |
| | | |||||
| * | [Tooling] Complete backend tooling and validation | Richard-Qin-X | 2026-08-10 | 90 | -2995/+11345 |
| | | |||||
| * | [MC] Implement RV64GC object emission | Richard-Qin-X | 2026-08-09 | 52 | -58/+7361 |
| | | |||||
| * | [MC] Implement RV64GC assembly emission | Richard-Qin-X | 2026-08-09 | 52 | -3/+5886 |
| | | |||||
| * | [CodeGen] Implement machine optimization and scheduling | Richard-Qin-X | 2026-08-09 | 91 | -40/+14212 |
| | | |||||
| * | [CodeGen] Implement RV64GC frame lowering | Richard-Qin-X | 2026-08-09 | 29 | -9/+4264 |
| | | |||||
| * | [CodeGen] Complete register allocation infrastructure | Richard-Qin-X | 2026-08-09 | 70 | -194/+7008 |
| | | |||||
| * | [CodeGen] Implement machine liveness analysis | Richard-Qin-X | 2026-08-09 | 15 | -53/+3294 |
| | | |||||
| * | [CodeGen] Implement machine SSA exit | Richard-Qin-X | 2026-08-08 | 22 | -56/+3594 |
| | | |||||
| * | [CodeGen] Implement RV64GC legalization and instruction selection | Richard-Qin-X | 2026-08-08 | 11 | -0/+5793 |
| | | |||||
| * | [CodeGen] Complete machine pass and ABI lowering infrastructure | Richard-Qin-X | 2026-08-07 | 25 | -0/+10628 |
| | | |||||
| * | [RISCV] Complete RV64GC instruction description | Richard-Qin-X | 2026-08-07 | 11 | -87/+5460 |
| | | |||||
| * | [RISCV] Add generated RV64GC register description | Richard-Qin-X | 2026-08-07 | 19 | -9/+2062 |
| | | | | | Add an RCC-owned build-time target description generator and the initial RV64GC register model. Keep ABI aliases separate from canonical MIR register spelling. | ||||
| * | [MIR] Complete document parsing and printing | Richard-Qin-X | 2026-08-07 | 45 | -218/+11465 |
| | | | | | Implement the single LLVM-compatible MIR document path with structured\nembedded-IR references, YAML mappings, constants, and deterministic printing.\n\nKeep pointer LLT widths DataLayout-derived, reject layout-dependent output\nwhen embedded IR is omitted, and record the fixed-baseline compatibility\ncontract and regression coverage. | ||||
| * | [MIR] Add machine instruction parser and printer | Richard-Qin-X | 2026-08-04 | 15 | -19/+5394 |
| | | | | | Parse MachineFunction bodies in two passes and serialize the same MachineIR model in complete or readable form. Extend instruction attachments and memory operands required by the fixed LLVM MIR grammar, and cover deterministic parse-print round trips. | ||||
| * | [MIR] Add machine instruction lexer | Richard-Qin-X | 2026-08-03 | 5 | -4/+1569 |
| | | |||||
| * | [CodeGen] Add target MIR text schema | Richard-Qin-X | 2026-08-03 | 20 | -50/+1363 |
| | | |||||
| * | [CodeGen] Model complete MIR operands | Richard-Qin-X | 2026-08-03 | 15 | -5/+1572 |
| | | |||||
| * | [CodeGen] Model machine debug metadata | Richard-Qin-X | 2026-08-02 | 22 | -38/+814 |
| | | | | | Replace marker-based metadata operands with scoped references and inline DIExpression values. Add structured DILocation chains, instruction debug attachments, stack variable records, and entry values with unified verification, builder support, and deterministic cloning. | ||||
| * | [CodeGen] Add machine metadata graphs | Richard-Qin-X | 2026-08-02 | 8 | -0/+492 |
| | | | | | Model ordered function-local metadata tuples with explicit embedded-IR and MachineFunction reference scopes. Preserve slot zero, distinct nodes, forward references, and cyclic graphs while verifying function-local closure and cloning the graph deterministically. | ||||
| * | [CodeGen] Track debug instruction identities | Richard-Qin-X | 2026-08-01 | 10 | -0/+471 |
| | | | | | Add monotonic MIR debug instruction numbers and structured value substitutions. Verify live number uniqueness, meta-instruction exclusion, substitution graph integrity, and target subregister references while preserving debug identities during cloning. | ||||
| * | [CodeGen] Add machine function records | Richard-Qin-X | 2026-08-01 | 8 | -0/+724 |
| | | | | | Represent callee-saved registers, call-site information, called globals, and prefetch targets in the core MachineIR model. Verify target and instruction references and remap StableIDs during explicit function cloning. | ||||
| * | [CodeGen] Add MIR function and block attributes | Richard-Qin-X | 2026-08-01 | 15 | -26/+745 |
| | | |||||
| * | [MIR] Add compilation unit reference context | Richard-Qin-X | 2026-08-01 | 12 | -0/+1251 |
| | | |||||
| * | [Support] Add MIR YAML parsing infrastructure | Richard-Qin-X | 2026-08-01 | 8 | -5/+1890 |
| | | | | | Introduce source ranges, deterministic text diagnostics, and an independently authored YAML stream parser for the fixed LLVM MIR container shapes. Record the compatibility matrix and Stage 6 validation state. | ||||
| * | [CodeGen] Add target register descriptions | Richard-Qin-X | 2026-08-01 | 24 | -156/+2821 |
| | | | | | Introduce an immutable register topology for units, subregisters, classes, banks, and preserved masks. Bind instruction descriptors to this topology through TargetDescription so builders, cloning, and verification share one target-aware contract. Add the test-only TestTarget model and positive and negative target fact coverage. | ||||
