Before this patch, running the x64 new backend would require both
compiling with --features experimental_x64 and running with
`use_new_backend`.
This patches changes this behavior so that the runtime flag is not
needed anymore: using the feature flag will enforce usage of the new
backend everywhere, making using and testing it much simpler:
cargo run --features experimental_x64 ;; other CLI options/flags
This also gives a hint at what the meta language generation would look
like after switching to the new backend.
Compiling only with the x64 codegen flag gives a nice compile time speedup.
This commit adds support for generating stackmaps at safepoints to the
new backend framework and to the AArch64 backend in particular. It has
been tested to work with SpiderMonkey.
This commit adds the inital support to allow reftypes to flow through
the program when targetting aarch64. It also adds a fix to the
`ModuleTranslationState` needed to send R32/R64 types over from the
SpiderMonkey embedding.
This commit does not include any support for safepoints in aarch64
or the `MachInst` infrastructure; that is in the next commit.
This commit also makes a drive-by improvement to `Bint`, avoiding an
unneeded zero-extension op when the extended value comes directly from a
conditional-set (which produces a full-width 0 or 1).
The main issue with the InstSize enum was that it was used both for
GPR and SIMD & FP operands, even though machine instructions do not
mix them in general (as in a destination register is either a GPR
or not). As a result it had methods such as sf_bit() that made
sense only for one type of operand.
Another issue was that the enum name was not reflecting its purpose
accurately - it was meant to represent an instruction operand size,
not an instruction size, which is fixed in A64 (always 4 bytes).
Now the enum is split into one for GPR operands and another for
scalar SIMD & FP operands.
Copyright (c) 2020, Arm Limited.
* Switch CI back to nightly channel
I think all upstream issues are now fixed so we should be good to switch
back to nightly from our previously pinned version.
* Fix doc warnings
This commit is intended to update wasmparser to 0.59.0. This primarily
includes bytecodealliance/wasm-tools#40 which is a large update to how
parsing and validation works. The impact on Wasmtime is pretty small at
this time, but over time I'd like to refactor the internals here to lean
more heavily on that upstream wasmparser refactoring.
For now, though, the intention is to get on the train of wasmparser's
latest `main` branch to ensure we get bug fixes and such.
As part of this update a few other crates and such were updated. This is
primarily to handle the new encoding of `ref.is_null` where the type is
not part of the instruction encoding any more.
We were accidentally always running the `fib-debug/main` example because of
shenanigans with alphabetical ordering and hard coding "main.exe" as the command
we run. Now we properly detect which example we built and run the appropriate
executable.
The canonical examples tests will be the examples run via `cargo run -p
run-examples` from now on. Those tests have two advantages over the ones being
deleted:
1. They will automatically pick up new examples and run/test them.
2. They support all of Linux, MacOS, and Windows, while the tests being deleted
are Linux only.
This required that `wasm_val_t` have a `Drop` implementation, an explicit
`Clone` implementation, and no longer be `Copy`, which rippled out through the
crate a bit.
Additionally, `wasm_func_call` and friends were creating references to
uninitialized data for its out pointers and assigning to them. As soon as
`wasm_val_t` gained a `Drop` impl and tried to drop the old value of the
assignment (which is uninitialized data), then things blew up. The fix is to
properly represent the out pointers with `MaybeUninit`, and use `ptr::write` to
initialize them without dropping the old data.
Part of #929
The ARM book says that the immr field should contain (-count % 64); the
existing code was approximating this with (64 - count), which is not
correct for a zero count.
Alex Crichton
MaxGraey
Benjamin Bouvier
Anton Kirilov
Chris Fallin
Joey Gouly
bjorn3
Nick Fitzgerald
Peter Huene
Yury Delendik
whitequark
Jakub Konka
Rochet2
Andrew Brown