Andrea Gussoni b10024bfc8 InlineHelpers: limit loop-unroll to x86
Only the x86 softfloat helpers iterate over the vector-register lanes
with bounded `for` loops (e.g. the packed `helper_*p[sd]` operate on
`XMM/ZMM_D(i)`). `fix-helpers` can only rewrite the per-lane CSV
accesses once those loops are unrolled. This step is not needed on other
archs, so we avoid it to save space on the generated libtcg helpers and
build time.
2026-06-19 15:35:18 +02:00
2024-11-21 15:54:23 +01:00
2026-06-19 15:35:18 +02:00
2026-06-15 11:47:41 +02:00
2026-06-19 15:35:18 +02:00
2026-06-19 11:07:44 +02:00
2026-06-19 15:35:18 +02:00
2025-09-10 12:05:15 +02:00
2024-02-26 16:02:52 +01:00
2026-06-15 17:28:22 +02:00
2024-11-21 15:54:23 +01:00

The rev.ng binary analysis framework and decompiler

Welcome to the rev.ng project!

rev.ng is an open source binary analysis framework based on LLVM and QEMU. It features a full-fledged decompiler. It's developed by rev.ng Labs in the open.

Getting started

Head to the online documentation to get started (source).

Getting in touch

You can get in touch with us through the following means:

  • X/Twitter, for the latest news about the project;
  • Discord, a place for chatting for users and developers;
  • Discourse, for user discussions;
  • GitHub, for development issues and PRs;
  • E-mail, for anything else;

License

The rev.ng binary analysis framework and decompiler is Free Software.

rev.ng depends on certain on GPLv2 components (in particular, QEMU), so the project is GPLv2-licensed as a whole. However, the individual files are licensed under the MIT license.

S
Description
Automated archival mirror of github.com/revng/revng
Readme GPL-2.0 24 MiB
Languages
C++ 77.1%
Python 12.6%
MLIR 4.4%
CMake 1.9%
LLVM 1.4%
Other 2.5%