This commit renames `wasi_config_set_std[in|out|err]` to
`wasi_config_set_std[in|out|err]_file` so we can reserve the former for
when the C API supports a stream abstraction.
This commit makes `WasiCtxBuilder` take `&mut Self` and return `&mut
Self` for its methods. This is needed to allow for the same
(unmoved) `WasiCtxBuilder` to be used when building a WASI context.
Also fixes up the C API to remove the unnecessary `Box::from_raw` and
`forget` calls which were previously needed for the moving version of
`WasiCtxBuilder`.
This commit implements an initial WASI C API that can be used to instantiate
and configure a WASI instance from C.
This also implements a `WasiBuilder` for the C# API enabling .NET hosts to bind
to Wasmtime's WASI implementation.
This commit does a bit of everything: refactors bits here and there,
fixes a bug discovered in another #701, and combines all structs that
we used in `yanix` and `wasi-common` crates to represent file types
on *nix into one struct, `yanix::file::FileType`.
Up until now, in `yanix`, we've had two separate structs used to
represent file types on the host: `yanix::dir::FileType` and
`yanix::file::SFlags` (well, not quite, but that was its main use).
They both were used in different context (the former when parsing
`dirent` struct, and the latter when parsing `stat` struct), they
were C-compatible (as far as their representation goes), and as it
turns out, they shared possible enumeration values. This commit
combines them both into an idiomatic Rust enum with the caveat that
it is now *not* C-compatible, however, I couldn't find a single use
where that would actually matter, and even if it does in the future,
we can simply add appropriate impl methods.
The combine `yanix::file::FileType` struct can be constructed in two
ways: 1) either from `stat.st_mode` value (and while we're here,
now it's done correctly according to POSIX which fixes the bug mentioned
in VFS impl PR #701), or 2) from `dirent.d_type` value. Also, since we now
have one struct for representing both contexts, this cleans up nicely
a lot of duplicated code in `host` module.
This allows getN to return a detailed explanation of any type signature
mismatch, and makes it easy to just use `?` on the result of getN rather
than constructing a (necessarily vaguer) error message in the caller.
* Cargo.lock: Update, to no longer use multiple versions of autocfg
* Update wasmtime-debug and wasmtime-profiling to current gimli 0.20.0
This also eliminates duplicate versions of gimli and arrayvec, and
eliminates the nodrop dependency entirely.
* Update wasmtime-profiling to goblin 0.1.3 and object 0.17.0
This also eliminates two duplicate versions of goblin, and duplicate
versions of proc-macro2, quote, syn, scroll_derive, and unicode-xid.
* Update wasmtime-profiling to current scroll 0.10.1
This eliminates duplicate versions of scroll.
* Update wasmtime-profiling to current target-lexicon 0.10.0
This eliminates duplicate versions of target-lexicon.
* Update wasmtime-interface-types to current walrus and wasm-webidl-bindings
This also eliminates the oldest of the three duplicate versions of
wasmparser.
* Update wasmtime-wast to current wast 8.0.0
This eliminates one of the duplicate versions of wast.
* Func: Number type arguments rather than using successive letters
This simplifies future extension, and avoids potential conflicts with
other type argument names.
* Extend Func::getN up to get10, allowing up to 10-argument functions
Patch adds support for the perf jitdump file specification.
With this patch it should be possible to see profile data for code
generated and maped at runtime. Specifically the patch adds support
for the JIT_CODE_LOAD and the JIT_DEBUG_INFO record as described in
the specification. Dumping jitfiles is enabled with the --jitdump
flag. When the -g flag is also used there is an attempt to dump file
and line number information where this option would be most useful
when the WASM file already includes DWARF debug information.
The generation of the jitdump files has been tested on only a few wasm
files. This patch is expected to be useful/serviceable where currently
there is no means for jit profiling, but future patches may benefit
line mapping and add support for additional jitdump record types.
Usage Example:
Record
sudo perf record -k 1 -e instructions:u target/debug/wasmtime -g
--jitdump test.wasm
Combine
sudo perf inject -v -j -i perf.data -o perf.jit.data
Report
sudo perf report -i perf.jit.data -F+period,srcline
We've got some OOM fuzz test cases getting reported, but these aren't
very interesting. The OOMs, after some investigation, are confirmed to
be happening because the test is simply allocating thousands of
instances with massive tables, quickly exceeding the 2GB memory
threshold for fuzzing. This isn't really interesting because this is
expected behavior if you instantiate these sorts of modules.
This commit updates the fuzz test case generator to have a "prediction"
for each module how much memory it will take to instantiate it. This
prediction is then used to avoid instantiating new modules if we predict
that it will exceed our memory limit. The limits here are intentionally
very squishy and imprecise. The goal here is to still generate lots of
interesting test cases, but not ones that simply exhaust memory
trivially.
* witx tagged unions: updates to wig to use new semantics
* wig: emit a `#variant: ()` union variant for empty variants
* wasi-common: translate to use tagged unions
* update to flattened layout of event struct
* wig: generate layout tests, and delete bindgen ones
the bindgen tests became out-of-date with the latest changes to the
representation of unions, and the re-jiggering of various struct
definitions that went along with it.
* wasi: point at master with tagged union PR merged
* fix event struct repr on windows
* Add API to statically assert signature of a `Func`
This commit add a family of APIs to `Func` named `getN` where `N` is the
number of arguments. Each function will attempt to statically assert the
signature of a `Func` and, if matching, returns a corresponding closure
which can be used to invoke the underlying function.
The purpose of this commit is to add a highly optimized way to enter a
wasm module, performing type checks up front and avoiding all the costs
of boxing and unboxing arguments within a `Val`. In general this should
be much more optimized than the previous `call` API for entering a wasm
module, if the signature is statically known.
* rustfmt
* Remove stray debugging
This commit shrinks the `RelocationTarget` enumeration to remove
intrinsic-related relocations since they are no longer used. Instead
these function calls are done indirectly via a table in the `VMContext`.
This means that all of this is essentially dead code!
* Fix a possible use-after-free with `Global`
This commit fixes an issue with the implementation of the
`wasmtime::Global` type where if it previously outlived the original
`Instance` it came from then you could run into a use-after-free. Now
the `Global` type holds onto its underlying `InstanceHandle` to ensure
it retains ownership of the underlying backing store of the global's
memory.
* rustfmt
Previously `Instance` was always allocated with `mmap`. This was done to
future-proof `Instance` for allowing storing the memory itself inline
with an `Instance` allocation, but this can actually be done with
`alloc`/`dealloc` since they take an alignment. By using `malloc`/`free`
we can avoid fragmentation as well as hook into standard leak tracking
mechanisms.
* Generate trampolines based on signatures
Instead of generating a trampoline-per-function generate a
trampoline-per-signature. This should hopefully greatly increase the
cache hit rate on trampolines within a module and avoid generating a
function-per-function.
* Update crates/runtime/src/traphandlers.rs
Co-Authored-By: Sergei Pepyakin <s.pepyakin@gmail.com>
Co-authored-by: Sergei Pepyakin <s.pepyakin@gmail.com>
* Update cranelift to 0.58.0
* Update `wasmprinter` dep to require 0.2.1
We already had it in the lock file, but this ensures we won't ever go back down.
* Ensure that our error messages match `assert_invalid`'s
The bulk of this work was done in
https://github.com/bytecodealliance/wasmparser/pull/186 but now we can test it
at the `wasmtime` level as well.
Fixes#492
* Stop feeling guilty about not matching `assert_malformed` messages
Remove the "TODO" and stop printing warning messages. These would just be busy
work to implement, and getting all the messages the exact same relies on using
the same structure as the spec interpreter's parser, which means that where you
have a helper function and they don't, then things go wrong, and vice versa. Not
worth it.
Fixes#492
* Enable (but ignore) the reference-types proposal tests
* Match test suite directly, instead of roundabout starts/endswith
* Enable (but ignore) bulk memory operations proposal test suite
* Add some debug logging to fuzzers
This is useful when trying to figure out what happened locally when
debugging fuzz test cases. By setting `RUST_LOG=wasmtime_fuzzing=debug`
you can get wasm files written to disk and for the API calls test case
see what API calls are being made.
* Also write out `*.wat` files
* rustfmt
* Remove return value from `log_wasm`
* Remove unused import
* Remove the `action` and `context` modules from `wasmtime_jit`
These modules are now no longer necessary with the `wasmtime` crate
fleshed out, and they're entirely subsumed by the `wasmtime` API as
well.
* Remove some more modules
* Remove the `jit_function_registry` global state
This commit removes on the final pieces of global state in wasmtime
today, the `jit_function_registry` module. The purpose of this module is
to help translate a native backtrace with native program counters into a
wasm backtrace with module names, function names, and wasm module
indices. To that end this module retained a global map of function
ranges to this metadata information for each compiled function.
It turns out that we already had a `NAMES` global in the `wasmtime`
crate for symbolicating backtrace addresses, so this commit moves that
global into its own file and restructures the internals to account for
program counter ranges as well. The general set of changes here are:
* Remove `jit_function_registry`
* Remove `NAMES`
* Create a new `frame_info` module which has a singleton global
registering compiled module's frame information.
* Update traps to use the `frame_info` module to symbolicate pcs,
directly extracting a `FrameInfo` from the module.
* Register and unregister information on a module level instead of on a
per-function level (at least in terms of locking granluarity).
This commit leaves the new `FRAME_INFO` global variable as the only
remaining "critical" global variable in `wasmtime`, which only exists
due to the API of `Trap` where it doesn't take in any extra context when
capturing a stack trace through which we could hang off frame
information. I'm thinking though that this is ok, and we can always
tweak the API of `Trap` in the future if necessary if we truly need to
accomodate this.
* Remove a lazy_static dep
* Add some comments and restructure
* Add more CLI flags for wasm features
This commit adds a few more flags to enable wasm features via the CLI,
mirroring the existing `--enable-simd` flag:
* `--enable-reference-types`
* `--enable-multi-value`
* `--enable-threads`
* `--enable-bulk-memory`
Additionally the bulk memory feature is now automatically enabled if
`reference-types` or `threads` are enabled since those two proposals
largely depend on `bulk-memory`.
* Add --enable-all to enable all wasm features
* Update src/lib.rs
Co-Authored-By: Peter Huene <peterhuene@protonmail.com>
* Apply suggestions from code review
Co-Authored-By: Peter Huene <peterhuene@protonmail.com>
Co-authored-by: Peter Huene <peterhuene@protonmail.com>
* Fuzz the multi-value support
This commit enables multi-value by default for the fuzzers, in theory
allowing us to find panics and such in the multi-value implementation.
Or even runtime errors through the differential fuzzing!
* Don't fuzz differential on multi value
There's not really much reason to amortize the cost of this mtime
calculation here since it's only done with debug assertions anyway, so
let's avoid an extra dependency and just have a function do it inline.
* Remove all global state from the caching system
This commit is a continuation of an effort to remove usages of
`lazy_static!` and similar global state macros which can otherwise be
accomodated with passing objects around. Previously there was a global
cache system initialized per-process, but it was initialized in a bit of
a roundabout way and wasn't actually reachable from the `wasmtime` crate
itself. The changes here remove all global state, refactor many of the
internals in the cache system, and makes configuration possible through
the `wasmtime` crate.
Specifically some changes here are:
* Usage of `lazy_static!` and many `static` items in the cache module
have all been removed.
* Global `cache_config()`, `worker()`, and `init()` functions have all
been removed. Instead a `CacheConfig` is a "root object" which
internally owns its worker and passing around the `CacheConfig` is
required for cache usage.
* The `wasmtime::Config` structure has grown options to load and parse
cache files at runtime. Currently only loading files is supported,
although we can likely eventually support programmatically configuring
APIs as well.
* Usage of the `spin` crate has been removed and the dependency is removed.
* The internal `errors` field of `CacheConfig` is removed, instead
changing all relevant methods to return a `Result<()>` instead of
storing errors internally.
* Tests have all been updated with the new interfaces and APIs.
Functionally no real change is intended here. Usage of the `wasmtime`
CLI, for example, should still enable the cache by default.
* Fix lightbeam compilation
* Remove global state for trap registration
There's a number of changes brought about in this commit, motivated by a
few things. One motivation was to remove an instance of using
`lazy_static!` in an effort to remove global state and encapsulate it
wherever possible. A second motivation came when investigating a
slowly-compiling wasm module (a bit too slowly) where a good chunk of
time was spent in managing trap registrations.
The specific change made here is that `TrapRegistry` is now stored
inside of a `Compiler` instead of inside a global. Additionally traps
are "bulk registered" for a module rather than one-by-one. This form of
bulk-registration allows optimizing the locks used here, where a lock is
only held for a module at-a-time instead of once-per-function.
With these changes the "unregister" logic has also been tweaked a bit
here and there to continue to work. As a nice side effect the `Compiler`
type now has one fewer field that requires actual mutability and has
been updated for multi-threaded compilation, nudging us closer to a
world where we can support multi-threaded compilation. Yay!
In terms of performance improvements, a local wasm test file that
previously took 3 seconds to compile is now 10% faster to compile,
taking ~2.7 seconds now.
* Perform trap resolution after unwinding
This avoids taking locks in signal handlers which feels a bit iffy...
* Remove `TrapRegistration::dummy()`
Avoid an case where you're trying to lookup trap information from a
dummy module for something that happened in a different module.
* Tweak some comments
The intention of the `wasmtime` crate was to disable this verifier by
default, but it looks like cranelift actually has it turned on by
default which was making our documentation incorrect!
This was discovered by seeing a number of timeouts when fuzzing. The
debug verifier is great for fuzzing, however, so fuzzing is updated to
enable this unconditionally, meaning we'll still have timeouts. For
general users though this should make the documentation correct that the
`wasmtime` crate, by default, disables the debug verifier.
Investigating a susprisingly slow-compiling module recently, it turns
out that if you create a wasm module with 40k empty functions (e.g.
`(module (func) (func) (func) ...)`) then it takes **3 seconds** to
compile and drop via the CLI locally on a Linux system. This seems like
an extraordinary amount of time for "doing nothing", and after some
profiling I found that basically all of the time was spent in
`__deregister_frame` calls.
Poking around in the source it looks like libgcc is managing some form
of linked list, and by deregistering in the LIFO order instead of FIFO
order it avoids a quadratic search of all registered functions. Now that
being said it's still pretty bad to do a linear search all the time, and
nothing will be fixed if there are *two* instances both with 40k
functions.
For now though I hope that this will patch over the performance issue
and we can figure out better ways to manage this in the future.
* Reimplement `wasmtime-wasi` on top of `wasmtime`
This commit reimplements the `wasmtime-wasi` crate on top of the
`wasmtime` API crate, instead of being placed on top of the `wasmtime-*`
family of internal crates. The purpose here is to continue to exercise
the API as well as avoid usage of internals wherever possible and
instead use the safe API as much as possible.
The `wasmtime-wasi` crate's API has been updated as part of this PR as
well. The general outline of it is now:
* Each module snapshot has a `WasiCtxBuilder`, `WasiCtx`, and `Wasi`
type.
* The `WasiCtx*` types are reexported from `wasi-common`.
* The `Wasi` type is synthesized by the `wig` crate's procedural macro
* The `Wasi` type exposes one constructor which takes a `Store` and a
`WasiCtx`, and produces a `Wasi`
* Each `Wasi` struct fields for all the exported functions in that wasi
module. They're all public an they all have type `wasmtime::Func`
* The `Wasi` type has a `get_export` method to fetch an struct field by
name.
The intention here is that we can continue to make progress on #727 by
integrating WASI construction into the `Instance::new` experience, but
it requires everything to be part of the same system!
The main oddity required by the `wasmtime-wasi` crate is that it needs
access to the caller's `memory` export, if any. This is currently done
with a bit of a hack and is expected to go away once interface types are
more fully baked in.
* Remove now no-longer-necessary APIs from `wasmtime`
* rustfmt
* Rename to from_abi
