xgadget

crates.io GitHub Actions

Fast, parallel, cross-variant ROP/JOP gadget search for x86 (32-bit) and x64 (64-bit) binaries. Uses the iced-x86 disassembler library.

Current state: decent test coverage, but still in beta. Issues/PRs welcome :)

About

To the best of my knowledge, xgadget is the first gadget search tool to have these features:

Other features include:

API Usage

Find gadgets:

```rust use xgadget;

let maxgadgetlen = 5; let search_config = xgadget::SearchConfig::DEFAULT;

// Search single binary let bin1 = xgadget::Binary::frompathstr("/path/to/binv1").unwrap(); let bins = vec![bin1]; let gadgets = xgadget::findgadgets(&bins, maxgadgetlen, searchconfig).unwrap(); let stackpivotgadgets = xgadget::filterstack_pivot(&gadgets);

// Search for cross-variant gadgets let bin1 = xgadget::Binary::frompathstr("/path/to/binv1").unwrap(); let bin2 = xgadget::Binary::frompathstr("/path/to/binv2").unwrap(); let bins = vec![bin1, bin2]; let crossgadgets = xgadget::findgadgets(&bins, maxgadgetlen, searchconfig).unwrap(); let crossregwritegadgets = xgadget::filterstacksetregs(&crossgadgets); ```

CLI Usage

Run xgadget --help:

``` xgadget v0.4.0

About: Fast, parallel, cross-variant ROP/JOP gadget search for x86/x64 binaries. Cores: 8 logical, 8 physical

USAGE: xgadget [FLAGS] [OPTIONS]

FLAGS: -t, --att Display gadgets using AT&T syntax [default: Intel syntax] -c, --check-sec Run checksec on the 1+ binaries instead of gadget search -d, --dispatcher Filter to potential JOP 'dispatcher' gadgets [default: all gadgets] -e, --extended-fmt Print in terminal-wide format [default: only used for partial match search] -h, --help Prints help information --inc-call Include gadgets containing a call [default: don't include] --inc-imm16 Include '{ret, ret far} imm16' (e.g. add to stack ptr) [default: don't include] -j, --jop Search for JOP gadgets only [default: ROP, JOP, and SYSCALL] -n, --no-color Don't color output, useful for UNIX piping [default: color output] -m, --partial-match Include cross-variant partial matches [default: full matches only] -w, --reg-write Filter to 'pop {reg} * 1+, {ret or ctrl-ed jmp/call}' gadgets [default: all gadgets] -r, --rop Search for ROP gadgets only [default: ROP, JOP, and SYSCALL] -p, --stack-pivot Filter to gadgets that write the stack ptr [default: all gadgets] -s, --sys Search for SYSCALL gadgets only [default: ROP, JOP, and SYSCALL] -V, --version Prints version information

OPTIONS: -a, --arch For raw (no header) files: specify arch ('x8086', 'x86', or 'x64') [default: x64] -b, --bad-bytes Gadgets up to LEN instrs long. If 0: all gadgets, any length [default: 5] -f, --regex-filter Filter to gadgets matching a regular expression

ARGS:

CLI Build and Install (Recommended)

Build a dynamically-linked binary from source and install it locally:

bash cargo install xgadget --features cli-bin # Build on host (pre-req: https://www.rust-lang.org/tools/install)

CLI Binary Releases for Linux

Commits to this repo's master branch automatically run integration tests and build a statically-linked binary for 64-bit Linux. You can download it here and use the CLI immediately, instead of building from source. Static binaries for Windows may also be supported in the future.

The statically-linked binary is about 8x slower, presumably due to the built-in memory allocator for target x86_64-unknown-linux-musl. Building a dynamically-linked binary from source with the above cargo install command is highly recommended.

~~Yeah, but can it do 10 OS kernels under 10 seconds?!~~ Repeatable Benchmark Harness

bash bash ./benches/bench_setup_ubuntu.sh # Ubuntu-specific, download/build 10 kernel versions cargo bench # Grab a coffee, this'll take a while...

On an i7-9700K (8C/8T, 3.6GHz base, 4.9 GHz max) machine with gcc version 8.4.0: the average runtime, to process all ten 54MB kernels simultaneously with a max gadget length of 5 instructions and full-match search for all gadget types (ROP, JOP, and syscall gadgets), is only 5.8 seconds! Including partial matches as well takes just 7.2 seconds.

Acknowledgements

This project started as an optimized solution to Chapter 8, exercise 3 of "Practical Binary Analysis" by Dennis Andreisse [6], and builds on the design outlined therein.

References