Reverse-Engineering Instruction Encodings (2001)
Web page · https://usenix.org/legacy/events/usenix01/full_papers/hsieh/hsieh_html
First seen by Oriel on .
AI summary
DERIVE reverse-engineers machine instruction encodings by assembling generated instruction variants and analyzing their machine code, producing C declarations for binary tools.
- Users specify instruction names, assembly syntax, and operand types; they do not provide bit offsets, widths, opcode values, or register encodings.
- Separate solvers infer register, immediate, and jump-target fields from assembler output, using an opcode mask and field descriptions.
- The tool solves all of MIPS, SPARC, Alpha, and PowerPC, plus most ARM and x86 instructions; unsupported cases include some non-contiguous fields and complex immediate transformations.
- A code-emitter generator converts DERIVE output into C macros; using them reduced the Kaffe JIT backend description by 40%.
- The method assumes stateless assembly and largely independent operand fields; it can also reveal differences between an architecture manual and an assembler.
#instruction encoding #assemblers #binary tools #jit compilers
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Lobsters: 6, latest observation ; +5 over observed period since
Comments
Comment count reported by the source.
- Lobsters: 0, latest observation ; 0 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Lobsters | 1q1rc1 | https://www.usenix.org/legacy/events/usenix01/full_papers/hsieh/hsieh_html/ Read discussion |