| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
|
|
|
| |
Reverts #4889
The spec is unclear on this, and that PR moved us to do what V8 does. But
it sounds like we should clarify the spec to do things the other way, so this
goes back to that.
|
| |
|
|
|
|
|
|
|
|
| |
call.without.effects has a specific form, where the last parameter is a
function reference, and that function reference must have the right type
for the other parameters if called with them:
(call $call.without.effects
(..i32..)
(..f64..)
(..function reference, which takes params i32 and f64..)
|
| |
|
|
|
|
| |
This allows emscripten to move these helper functions from JS library
imports to native wasm exports.
See https://github.com/emscripten-core/emscripten/issues/7273
|
| |
|
|
|
|
| |
In BINARYEN_PASS_DEBUG=2 we save the module before each pass, and if
validation fails afterwards, we print the module before. This PR does the same for
function-parallel passes - in that case, we can actually show the specific function
that broke validation, as opposed to the whole module.
|
| |
|
| |
I was reading these tests and failing to find the names script.
|
| |
|
|
|
| |
Also, add support for the `--binaryen-bin` flag to
`scripts/port_passes_tests_to_lit.py`. This is needed for folks who
don't do in-tree builds.
|
| |
|
|
| |
So it lives alongside `update_lit_checks.py` and `update_help_checks.py`
|
| |
|
|
| |
`pop`s type should be a supertype, not a subtype, of the tag's type
within `catch`.
|
| |
|
|
|
|
|
| |
* better
* fix
* undo
|
| |
|
| |
Like the 8-bit array variants, it takes 3 parameters.
|
| |
|
|
|
|
|
| |
For now this index is always 0, but we must emit it.
Also clean up the wat test a little - we don't have validation yet, but we should
not validate without a memory in that file.
|
| |
|
|
|
|
| |
This starts to matter with strings, it turns out. This change should make us
runnable in v8.
Spec: https://github.com/WebAssembly/gc/blob/main/proposals/gc/MVP.md#instructions-1
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
call.without.effects will turn into a normal call of the last parameter later,
(call $call.without.effects
A
B
(ref.func $foo)
)
;; => intrinsic lowering
(call $foo
A
B
)
SignaturePruning needs to be aware of that: we can't remove a parameter from $foo without
also updating relevant calls to $call.without.effects. Rather than handle that, just skip such
cases, and leave them to be optimized after intrinsics are lowered away.
|
| |
|
|
|
|
|
| |
A function literal (ref.func) should never reach a struct or array get, but
if there is a cast then it can look like they might arrive. We filter in ref.cast
which avoids that (since casting a function to a data type will trap), but
there is also br_on_cast which is not yet optimized. This PR adds code
to avoid an assert in readFromData in that case.
|
| |
|
|
|
|
| |
eqz(eqz(i32(x))) -> i32(x) != 0
eqz(eqz(i64(x))) -> i64(x) != 0
Only when shrinkLevel == 0 (prefer speed over binary size).
|
| |
|
|
| |
This is no longer needed by emscripten as of:
https://github.com/emscripten-core/emscripten/pull/16529
|
| |
|
|
|
|
|
|
| |
stripping data segments. (#4876)
This avoid a fatal crash in `--post-emscripten` where it tries to remove
data that is no longer part of the file.
This fixes bug introduced by #4871 that causes emscripten tests to fail.
|
| |
|
|
|
|
|
| |
RTTs were removed from the GC spec and if they are added back in in the future,
they will be heap types rather than value types as in our implementation.
Updating our implementation to have RTTs be heap types would have been more work
than deleting them for questionable benefit since we don't know how long it will
be before they are specced again.
|
| |
|
|
| |
Rather than doing it as a side effect of dumping the metadata in
wasm-emscripten-finalize.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
`--debug`. NFC (#4874)
For example I found it useful to able to do something like this:
```
$ BINARYEN_DEBUG=post-emscripten ./test/runner sometest
```
|
| |
|
|
|
|
|
|
|
| |
Sometimes the fuzzer produces valid modules that trap during instantiation. When
that happens, the JS harness used to run the fuzzer output in d8 would
previously throw an error, creating spurious fuzzer failures on valid modules.
Update fuzz_shell.js to catch and supress errors during instantiation (but not
validation) to avoid these spurious failures.
Fixes #4865.
|
| |
|
| |
Introduces the necessary APIs to use the type builder from C. Enables construction of compound heap types (arrays, structs and signatures) that may be recursive, including assigning concrete names to the built types and, in case of structs, their fields.
|
| | |
|
| |
|
|
|
|
|
|
| |
We already remove `__start_em_asm` and `__stop_em_asm`. This change
is needed since I want to start exporting `__start_em_js` and
`__stop_em_js` from emscripten without causing regressions.
As a followup I'm planning on moving all of the em_js and em_asm
stripping code it PostEmscripten.cpp.
|
| |
|
|
|
| |
It has been removed from the typed function references proposal, so we no longer
need to support it. Maintaining the test for `let` was difficult because
Binaryen could not emit either text or binary that actually used it.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
+ Move these rules to separate function;
+ Refactor them to use matches;
+ Add comments;
+ Handle rotational shifts as well;
+ Handle overflows for `<<`, `>>`, `>>>` shifts;
+ Add mixed rotate rules:
```rust
rotl(rotr(x, C1), C2) => rotr(x, C1 - C2)
rotr(rotl(x, C1), C2) => rotl(x, C1 - C2)
```
|
| |
|
| |
Avoids a "may fall through" warning.
|
| | |
|
| |
|
|
|
|
| |
We already require non-null literals to have non-null types, but with this
change we can enforce that constraint by construction. Also remove the default
behavior of creating a function reference literal with heap type `func`, since
there is always a more specific function type to use.
|
| |
|
| |
This can give us some chance to catch bugs like #4839 in the fuzzer.
|
| |
|
|
|
|
|
|
| |
Like RemoveUnusedModuleElements, places that build graphs of function
reachability must special-case the call-without-effects intrinsic. Without that,
it looks like a call to an import. Normally a call to an import is fine - it makes us
be super-pessimistic, as we think things escape all the way out - but in GC
for now we are assuming a closed world, and so we end up broken. To fix that,
properly handle the intrinsic case.
|
| |
|
|
|
|
| |
It was wasted work to see a drop and then check if we can replace it with
a drop of its child, which is identical to the original state. This didn't cause
any harm (we'd not reduce code size, and stop eventually) but it did slow us
down.
|
| |
|
|
|
| |
When building Binaryen to wasm in debug mode,
1) Avoid post-link Binaryen changes with WASM_BIGINT and by avoiding O3
2) Don't use SINGLE_FILE, to make it easier to analyze the resulting wasm
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Refactor everywhere from:
```c++
for (size_t i = 0; i < indent; i++) {
o << ' ';
}
```
to:
```c++
o << std::string(indent, ' ');
```
### Motivation
It is much simpler and should produce smaller code.See godbolt:
https://godbolt.org/z/KMYMdn7z5
|
| |
|
| |
Make the C API match the JS API and fix an old bug where extra newlines were emitted.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
The stubs let precompute skip over them without erroring. With this PR we can
run the optimizer on strings code. We still can't run --fuzz-exec though, so we
can't run the fuzzer.
Also simplify the error strings in the earlier part of the file. All other code just
has "unimp" so we might as well do the same and not mention full names
there.
|
| |
|
|
|
|
|
| |
The previous code assumes if `last`'s type is unreachable it traps. But
it's not always the case because it can be other instructions like `br`
whose type is unreachable but doesn't necessarily trap.
Context:
https://github.com/WebAssembly/binaryen/pull/4827#discussion_r929395477
|
| |
|
|
|
| |
This lets wasm-reduce --enable-FOO work. Usually this is not needed as we do
enable all features by default, but sometimes it is nice to disable features (e.g. to
avoid reducing into a testcase that uses something the original wasm did not use).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There are two new potential problems that `GlobalTypeRewriter` can run into when
working with isorecursive types instead of nominal types. First, the refined
types may have replaced generic references with references to specific other
types, potentially creating new recursions and making the existing recursion
groups insufficient. Second, distinct types may be refined to structurally
identical types and those distinct input types may map the same output type,
potentially changing cast behavior.
Both of these problems are solved by putting all the new types in a single large
recursion group.
We do not currently account for the fact that types may be used in the external
interface of the module, but when we do, externalized types will be excluded
from optimizations and will not be affected by the creation of this single large
rec group.
Fixes #4816.
|
| |
|
|
|
|
|
|
|
|
|
| |
* Changing ref maps in wasm-binary to use a value of a vector of Name*
* clang-format
* Update src/wasm/wasm-binary.cpp
Co-authored-by: Thomas Lively <7121787+tlively@users.noreply.github.com>
Co-authored-by: Thomas Lively <7121787+tlively@users.noreply.github.com>
|
| | |
|
| |
|
|
|
|
| |
constants on RHS (#4808)
(x * C1) << C2 -> x * (C1 << C2)
(x << C1) * C2 -> x * (C2 << C1)
|
| |
|
|
|
|
|
|
|
| |
Add support for emitting the string type reference shorthands, which had
previously been omitted accidentally due to the `default` case in that switch.
Also avoid emitting shorthands for non-nullable reference types as a first step
towards transitioning the shorthands to represent nullable types instead. Not
emitting these shorthands at all will give V8 the flexibility it needs to change
its interpretation of the shorthands without breaking any workflows using
Binaryen.
|
