Merge with upstream #127
Merged
Merge with upstream #127
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Also cause it to return an error if the input is a core wasm module with zero `component-type` custom sections. Closes bytecodealliance#1673
This change attempts to add tests that check that shared-everything-threads features indeed fail when the proposal is not enabled. In doing this, I realized that a few features fail on other features first (GC, reftypes), which meant the tests couldn't be as exhaustive as I would have liked. This commit also changes the wording of one of the failures to be consistent across the proposal.
This change adopts the suggestion to use the "missing-features + X" functionality in `roundtrip.rs` to check more corner cases where shared-everything-threads must be enabled. It adds the reference-types proposal to the GC feature set to make this possible.
If an error was found before the code section a DWARF-related warning was printed but that's not relevant and is false information, so squash the warning.
* Update wasm-encoder's type encoding to match `wast` This updates it to not only match the structure but also binary output. This means that shorter forms are used a bit more and encodings for shared-everything-threads are additionally fixed. * Update wasm_encoder's reencoding of const exprs Use more generic helpers to support multiple instructions and additionally more instructions. This refactors the `parse_instruction` helper to return the instruction directly instead of unconditionally encoding into a `Function`. * Simplify some type encoding in `wast` No functional change here, just reducing some duplication of code. * Test wasm-encoder roundtrips of `roundtrip` tests This commit updates the wasm-roundtrip test suite to additionally run all binaries through the reencoding phase of `wasm-encoder` to assert the final binary matches `wast`. This is skipped for components for now as reencoding support isn't here for components just yet. * Fix error message when wasm binaries differ The condition here was inverted by accident meaning that more helpful error messages were not being generated.
* Update spec testsuite submodule to latest * Update all snapshots * Get all spec tests passing * Tweak some error messages and expected error messages. * Update validation of `try_table` to enable catching with `(ref exn)` * Update `json-from-wast` with remaining reference types * Update features enabled for `memory64` since it has rebased on the 3.0 spec.
…liance#1722) * Extend re-encode functionality to support components Signed-off-by: itowlson <ivan.towlson@fermyon.com> * WIP: indexes, reorg Signed-off-by: itowlson <ivan.towlson@fermyon.com> * Finish reencoding support for components to wasm-encoder This commit is an extension of bytecodealliance#1656 to support components in the reencoding support of wasm-encoder. The precise shape of the trait will probably change over time but this is at least a first stab at getting things working to map all the wasmparser types to wasm-encoder for components. All existing support for components in `reencode.rs` was moved to a new `ReencodeComponent` trait which inherits from the already-existing `Reencode` trait. This is all located in a new `component.rs` file beneath `reencode.rs` although crate-wise it's all showing up in the same `wasm_encoder::reencode` module. * Fix compat with Rust 1.78 --------- Signed-off-by: itowlson <ivan.towlson@fermyon.com> Co-authored-by: itowlson <ivan.towlson@fermyon.com>
This change fixes several `wast` typos and regularizes the encoding hex codes for the shared-everything-threads instructions to use lowercase like the rest of the file.
* threads: fix `ref.i31_shared` encoding In the initial implementation, this encoding was incorrectly emitted as `0xFE 0x1F`, clobbering the encoding for `i64.atomic.rmw.add` (probably a copy-paste error). This change updates `wasm-encoder` to use the correct `0xFE 0x72` encoding. * threads: fix `ref.i31_shared` stack type This instruction should return the `shared` version of `ref.i31`. * Add a simple re-encodable test
* threads: rename tests This change simply renames the shared-everything-threads tests to use more conventional names within the repository. * threads: add more Binaryen tests @tlively has been working on Binaryen support for shared-everything-threads and has upstreamed several [tests] that are useful here. This change incorporates some "bottom type" checks for arrays and structs as a `shared-ref_eq.wast` test checking `ref.eq` behavior for shared types (renamed to `ref-equivalence.wast` here). To make these added tests pass, validation for `ref.eq` now allows it to check both shared and ushared `eqref` types the the `ref.i31_shared` returns the correct type. [tests]: https://github.com/WebAssembly/binaryen/blob/main/test/spec [shared-ref_eq.wast]: https://github.com/WebAssembly/binaryen/blob/main/test/spec/shared-ref_eq.wast * threads: add polymorphism over `shared` Certain instructions accept both shared and unshared references: `ref.eq`, `i31.get_s`, `i31.get_u`, `array.len`, `any.convert_extern`, and `extern.convert_any`. Validation needs to handle these special cases to pass Binaryen's `polymorphism.wast` test. This change refactors the `pop_operand` family of functions with several helpers to add `pop_maybe_shared_ref`; `pop_maybe_shared_ref` lets the validator pass an unshared `RefType` expectation but adapts it to being shared if that is what is on the stack. * Undo previous refactoring; reuse `OperatorValidatorTemp::pop_ref` This change undoes previous work to rationalize the duplication in `_pop_operand`, `pop_ref`, and now `pop_maybe_shared_ref`. As suggested by @alexcrichton, this attempt keeps the status quo by reusing `pop_ref` and doing a bit more work after the fact (re-checking the subtype relationship). * Handle `ref.eq` types from unreachable code This validator's existing logic assumed that the bottom and heap bottom types matched any reference type (see `_pop_operand`). This change surfaces that logic to `visit_ref_eq` as well: if `pop_maybe_shared_ref` returns `None` (a bottom or heap bottom) for either value to `ref.eq`, we skip checking the shared-ness match, assuming like we do in `_pop_operand` that a bottom or heap bottom matches any reference type (we've already checked that the types are subtypes of `eqref`). * Add TODO for unshared propagation in unreachable code * Return shared `bool` from `pop_maybe_shared_ref` * Rename `is_shared_ref_type` to `is_shared`
…ance#1726) This was an idea I had after bytecodealliance#1719 landed to statically ensure that the asserts won't trip.
) * Add a new `wasm-tools component unbundle` command This commit adds a new subcommand to `wasm-tools component` named `unbundle` which is tasked with extracting core wasm modules out of a component. This is intended to showcase an example implementation of doing this and the main motivation is to extract core wasm modules out of a component for runtimes such as Wasmtime. This enables sharing machine code for WebAssembly modules between components that use the same core wasm module. This new subcommand is primarily built on `wasm_encoder::reencode` with a simple size-based heuristic for which modules to extract. * Test module output as well
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.