a2db70a8d0
- [x] `dlfcn.rs` - `dlopen()`-related code. - [x] `turbocall.rs` - Call trampoline JIT compiler. - [x] `repr.rs` - Pointer representation. Home of the UnsafePointerView ops. - [x] `symbol.rs` - Function symbol related code. - [x] `callback.rs` - Home of `Deno.UnsafeCallback` ops. - [x] `ir.rs` - Intermediate representation for values. Home of the `NativeValue` type. - [x] `call.rs` - Generic call ops. Home to everything related to calling FFI symbols. - [x] `static.rs` - static symbol support I find easier to work with this setup, I eventually want to expand TurboCall to unroll type conversion loop in generic calls, generate code for individual symbols (lazy function pointers), etc. |
||
|---|---|---|
| .. | ||
| 00_ffi.js | ||
| call.rs | ||
| callback.rs | ||
| Cargo.toml | ||
| dlfcn.rs | ||
| ir.rs | ||
| lib.rs | ||
| README.md | ||
| repr.rs | ||
| static.rs | ||
| symbol.rs | ||
| turbocall.rs | ||
deno_ffi
This crate implements dynamic library ffi.
Performance
Deno FFI calls have extremely low overhead (~1ns on M1 16GB RAM) and perform on par with native code. Deno leverages V8 fast api calls and JIT compiled bindings to achieve these high speeds.
Deno.dlopen generates an optimized and a fallback path. Optimized paths are
triggered when V8 decides to optimize the function, hence call through the Fast
API. Fallback paths handle types like function callbacks and implement proper
error handling for unexpected types, that is not supported in Fast calls.
Optimized calls enter a JIT compiled function "trampoline" that translates Fast
API values directly for symbol calls. JIT compilation itself is super fast,
thanks to tinycc. Currently, the optimized path is only supported on Linux and
MacOS.
To run benchmarks:
target/release/deno bench --allow-ffi --allow-read --unstable ./test_ffi/tests/bench.js