1532516a36
* Use relative `call` instructions between wasm functions
This commit is a relatively major change to the way that Wasmtime
generates code for Wasm modules and how functions call each other.
Prior to this commit all function calls between functions, even if they
were defined in the same module, were done indirectly through a
register. To implement this the backend would emit an absolute 8-byte
relocation near all function calls, load that address into a register,
and then call it. While this technique is simple to implement and easy
to get right, it has two primary downsides associated with it:
* Function calls are always indirect which means they are more difficult
to predict, resulting in worse performance.
* Generating a relocation-per-function call requires expensive
relocation resolution at module-load time, which can be a large
contributing factor to how long it takes to load a precompiled module.
To fix these issues, while also somewhat compromising on the previously
simple implementation technique, this commit switches wasm calls within
a module to using the `colocated` flag enabled in Cranelift-speak, which
basically means that a relative call instruction is used with a
relocation that's resolved relative to the pc of the call instruction
itself.
When switching the `colocated` flag to `true` this commit is also then
able to move much of the relocation resolution from `wasmtime_jit::link`
into `wasmtime_cranelift::obj` during object-construction time. This
frontloads all relocation work which means that there's actually no
relocations related to function calls in the final image, solving both
of our points above.
The main gotcha in implementing this technique is that there are
hardware limitations to relative function calls which mean we can't
simply blindly use them. AArch64, for example, can only go +/- 64 MB
from the `bl` instruction to the target, which means that if the
function we're calling is a greater distance away then we would fail to
resolve that relocation. On x86_64 the limits are +/- 2GB which are much
larger, but theoretically still feasible to hit. Consequently the main
increase in implementation complexity is fixing this issue.
This issue is actually already present in Cranelift itself, and is
internally one of the invariants handled by the `MachBuffer` type. When
generating a function relative jumps between basic blocks have similar
restrictions. This commit adds new methods for the `MachBackend` trait
and updates the implementation of `MachBuffer` to account for all these
new branches. Specifically the changes to `MachBuffer` are:
* For AAarch64 the `LabelUse::Branch26` value now supports veneers, and
AArch64 calls use this to resolve relocations.
* The `emit_island` function has been rewritten internally to handle
some cases which previously didn't come up before, such as:
* When emitting an island the deadline is now recalculated, where
previously it was always set to infinitely in the future. This was ok
prior since only a `Branch19` supported veneers and once it was
promoted no veneers were supported, so without multiple layers of
promotion the lack of a new deadline was ok.
* When emitting an island all pending fixups had veneers forced if
their branch target wasn't known yet. This was generally ok for
19-bit fixups since the only kind getting a veneer was a 19-bit
fixup, but with mixed kinds it's a bit odd to force veneers for a
26-bit fixup just because a nearby 19-bit fixup needed a veneer.
Instead fixups are now re-enqueued unless they're known to be
out-of-bounds. This may run the risk of generating more islands for
19-bit branches but it should also reduce the number of islands for
between-function calls.
* Otherwise the internal logic was tweaked to ideally be a bit more
simple, but that's a pretty subjective criteria in compilers...
I've added some simple testing of this for now. A synthetic compiler
option was create to simply add padded 0s between functions and test
cases implement various forms of calls that at least need veneers. A
test is also included for x86_64, but it is unfortunately pretty slow
because it requires generating 2GB of output. I'm hoping for now it's
not too bad, but we can disable the test if it's prohibitive and
otherwise just comment the necessary portions to be sure to run the
ignored test if these parts of the code have changed.
The final end-result of this commit is that for a large module I'm
working with the number of relocations dropped to zero, meaning that
nothing actually needs to be done to the text section when it's loaded
into memory (yay!). I haven't run final benchmarks yet but this is the
last remaining source of significant slowdown when loading modules,
after I land a number of other PRs both active and ones that I only have
locally for now.
* Fix arm32
* Review comments
Cranelift Code Generator
A Bytecode Alliance project
Cranelift is a low-level retargetable code generator. It translates a target-independent intermediate representation into executable machine code.
For more information, see the documentation.
For an example of how to use the JIT, see the JIT Demo, which implements a toy language.
For an example of how to use Cranelift to run WebAssembly code, see Wasmtime, which implements a standalone, embeddable, VM using Cranelift.
Status
Cranelift currently supports enough functionality to run a wide variety of programs, including all the functionality needed to execute WebAssembly MVP functions, although it needs to be used within an external WebAssembly embedding to be part of a complete WebAssembly implementation.
The x86-64 backend is currently the most complete and stable; other architectures are in various stages of development. Cranelift currently supports both the System V AMD64 ABI calling convention used on many platforms and the Windows x64 calling convention. The performance of code produced by Cranelift is not yet impressive, though we have plans to fix that.
The core codegen crates have minimal dependencies, support no_std mode (see below), and do not require any host floating-point support, and do not use callstack recursion.
Cranelift does not yet perform mitigations for Spectre or related security issues, though it may do so in the future. It does not currently make any security-relevant instruction timing guarantees. It has seen a fair amount of testing and fuzzing, although more work is needed before it would be ready for a production use case.
Cranelift's APIs are not yet stable.
Cranelift currently requires Rust 1.37 or later to build.
Contributing
If you're interested in contributing to Cranelift: thank you! We have a contributing guide which will help you getting involved in the Cranelift project.
Planned uses
Cranelift is designed to be a code generator for WebAssembly, but it is general enough to be useful elsewhere too. The initial planned uses that affected its design are:
- WebAssembly compiler for the SpiderMonkey engine in Firefox.
- Backend for the IonMonkey JavaScript JIT compiler in Firefox.
- Debug build backend for the Rust compiler.
- Wasmtime non-Web wasm engine.
Building Cranelift
Cranelift uses a conventional Cargo build process.
Cranelift consists of a collection of crates, and uses a Cargo
Workspace,
so for some cargo commands, such as cargo test, the --all is needed
to tell cargo to visit all of the crates.
test-all.sh at the top level is a script which runs all the cargo
tests and also performs code format, lint, and documentation checks.
Building with no_std
The following crates support `no_std`, although they do depend on liballoc:
- cranelift-entity
- cranelift-bforest
- cranelift-codegen
- cranelift-frontend
- cranelift-native
- cranelift-wasm
- cranelift-module
- cranelift-preopt
- cranelift
To use no_std mode, disable the std feature and enable the core feature. This currently requires nightly rust.
For example, to build `cranelift-codegen`:
cd cranelift-codegen
cargo build --no-default-features --features core
Or, when using cranelift-codegen as a dependency (in Cargo.toml):
[dependency.cranelift-codegen]
...
default-features = false
features = ["core"]
no_std support is currently "best effort". We won't try to break it, and we'll accept patches fixing problems, however we don't expect all developers to build and test no_std when submitting patches. Accordingly, the ./test-all.sh script does not test no_std.
There is a separate ./test-no_std.sh script that tests the no_std support in packages which support it.
It's important to note that cranelift still needs liballoc to compile. Thus, whatever environment is used must implement an allocator.
Also, to allow the use of HashMaps with no_std, an external crate called hashmap_core is pulled in (via the core feature). This is mostly the same as std::collections::HashMap, except that it doesn't have DOS protection. Just something to think about.
Log configuration
Cranelift uses the log crate to log messages at various levels. It doesn't
specify any maximal logging level, so embedders can choose what it should be;
however, this can have an impact of Cranelift's code size. You can use log
features to reduce the maximum logging level. For instance if you want to limit
the level of logging to warn messages and above in release mode:
[dependency.log]
...
features = ["release_max_level_warn"]
Editor Support
Editor support for working with Cranelift IR (clif) files: