| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| | |
|
| |
|
|
|
|
|
|
|
| |
Normally we use it when optimizing (above a certain level). This lets the user
prevent it from being used even then.
Also add optimization options to wasm-metadce so that this is possible
there as well and not just in wasm-opt (this also opens the door to running
more passes in metadce, which may be useful later).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Anything else right before an unreachable is removed by the main DCE
pass anyhow, but because of the structured form of BinaryenIR we can't remove
a drop. That is, this is the difference between
(i32.eqz
(i32.const 42)
(unreachable)
)
and
(drop
(call $foo)
)
(unreachable)
In both cases the unreachable is preceded by something we don't need,
but in the latter case it must remain in BinaryenIR for validation.
To optimize this, add a rule in StackIR.
Fixes #6715
|
| |
|
|
|
|
|
|
|
| |
Rename instructions `extern.internalize` into `any.convert_extern` and
`extern.externalize` into `extern.convert_any` to follow more closely
the spec. This was changed in
https://github.com/WebAssembly/gc/issues/432.
The legacy name is still accepted in text inputs and in the C and JS
APIs.
|
| |
|
|
|
|
|
|
|
| |
(#6709)
Previously the replacement select got the debug info, but we should also copy it
to the values, as often optimizations lead to one of those values remaining by
itself.
Similar to #6652 in general form.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RefTest (#6692)
CFP focuses on finding when a field always contains a constant, and then replaces
a struct.get with that constant. If we find there are two constant values, then in some
cases we can still optimize, if we have a way to pick between them. All we have is the
struct.get and its reference, so we must use a ref.test:
(struct.get $T x (..ref..))
=>
(select
(..constant1..)
(..constant2..)
(ref.test $U (..ref..))
)
This is valid if, of all the subtypes of $T, those that pass the test have
constant1 in that field, and those that fail the test have constant2. For
example, a simple case is where $T has two subtypes, $T is never created
itself, and each of the two subtypes has a different constant value.
This is a somewhat risky operation, as ref.test is not necessarily cheap.
To mitigate that, this is a new pass, --cfp-reftest that is not run by
default, and also we only optimize when we can use a ref.test on what
we think will be a final type (because ref.test on a final type can be
faster in VMs).
|
| |
|
|
|
| |
If an allocation does not escape, then we can compute ref.eq for it: when
compared to itself the result is 1, and when compared to anything else it
is 0 (since it did not escape, anything else must be different).
|
| |
|
| |
Add spec tests checking validation for structs and arrays.
|
| |
|
| |
Fixes #6690
|
| |
|
|
|
|
|
| |
This pass receives a list of functions to trace, and then wraps them in calls to
imports. This can be useful for tracing malloc/free calls, for example, but is
generic.
Fixes #6548
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(#6688)
If we have
(global $g (struct.new $S
(i32.const 1)
(struct.new $T ..)
(ref.func $f)
))
then before this PR if we wanted to read the middle field we'd stop, as it is non-constant.
However, we can un-nest it, making it constant:
(global $g.unnested (struct.new $T ..))
(global $g (struct.new $S
(i32.const 1)
(global.get $g.unnested)
(ref.func $f)
))
Now the field is a global.get of an immutable global, which is constant. Using this
technique we can handle anything in a struct field, constant or not. The cost of adding
a global is likely offset by the benefit of being able to refer to it directly, as that opens
up more opportunities later.
Concretely, this replaces the constant values we look for in GSI with a variant over
constants or expressions (we do still want to group constants, as multiple globals
with the same constant field can be treated as a whole). And we note cases where we
need to un-nest, and handle those at the end.
|
| |
|
|
|
|
|
|
|
|
|
| |
Implement binary and text parsing and printing of shared basic heap types and
incorporate them into the type hierarchy.
To avoid the massive amount of code duplication that would be necessary if we
were to add separate enum variants for each of the shared basic heap types, use
bit 0 to indicate whether the type is shared and replace `getBasic()` with
`getBasic(Unshared)`, which clears that bit. Update all the use sites to record
whether the original type was shared and produce shared or unshared output
without code duplication.
|
| |
|
|
| |
This is achieved by simply replacing the Literal with PossibleConstantValues, which
supports both Literals and Globals.
|
| |
|
|
|
|
|
| |
As an abbreviation, a `typeuse` can be given as just a list of parameters and
results, in which case it corresponds to the index of the first function type
with the same parameters and results. That function type must also be an MVP
function type, i.e. it cannot have a nontrivial rec group, be non-final, or have
a declared supertype. The parser did not previously implement all of these rules.
|
| |
|
|
| |
Also update the parser so that implicit type uses are not matched with shared
function types.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Parse the text format for shared composite types as described in the
shared-everything thread proposal. Update the parser to use 'comptype' instead
of 'strtype' to match the final GC spec and add the new syntactic class
'sharecomptype'.
Update the type canonicalization logic to take sharedness into account to avoid
merging shared and unshared types. Make the same change in the TypeMerging pass.
Ensure that shared and unshared types cannot be in a subtype relationship with
each other.
Follow-up PRs will add shared abstract heap types, binary parsing and emitting
for shared types, and fuzzer support for shared types.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We automatically copy debuginfo in replaceCurrent(), but there are a few
places that do other operations than simple replacements. call-utils.h will
turn a call_ref with a select target into two direct calls, and we were missing
the logic to copy debuginfo from the call_ref to the calls.
To make this work, refactor out the copying logic from wasm-traversal, into
debuginfo.h, and use it in call-utils.h.
debuginfo.h itself is renamed from debug.h (as now this needs to be included
from wasm-traversal, which nearly everything does, and it turns out some files
have internal stuff like a debug() helper that ends up conflicing with the old
debug namespace).
Also rename the old copyDebugInfo function to copyDebugInfoBetweenFunctions
which is more explicit. That is also moved from the header to a cpp file because
it depends on wasm-traversal (so we'd end up with recursive headers otherwise).
That is fine, as that method is called after copying a function, which is not that
frequent. The new copyDebugInfoToReplacement (which was refactored out of
wasm-traversal) is in the header because it can be called very frequently (every
single instruction we optimize) and we want it to get inlined.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We need StringLowering to modify even public types, as it must replace every
single stringref with externref, even if that modifies the ABI. To achieve that
we told it that all string-using types were private, which let TypeUpdater update
them, but the problem is that it moves all private types to a new single
rec group, which meant public and private types ended up in the same group.
As a result, a single public type would make it all public, preventing optimizations
and breaking things as in #6630 #6640.
Ideally TypeUpdater would modify public types while keeping them in the same
rec groups, but this may be a very specific issue for StringLowering, and that
might be a lot of work. Instead, just make StringLowering handle public types of
functions in a manual way, which is simple and should handle all cases that
matter in practice, at least in J2Wasm.
|
| |
|
|
|
|
|
|
|
|
| |
Without that logic we could end up dropping that particular effect. This actually
made a test pass when it should not: the modified test here has a function with
effects that are ok to remove, but it had a loop which adds MayNotReturn which
we should actually not remove, so it was removed erroneously.
To fix the test, add other effects there (local ones) that we can see are removable.
Also add a function with a loop to test that we do not remove an infinite loop,
which adds coverage for the fix here.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The old ordering in that pass did a topological sort while sorting by uses
both within topological groups and between them. That could be unoptimal
in some cases, however, and actually on J2CL output this pass made the
binary larger, which is how we noticed this.
The problem is that such a toplogical sort keeps topological groups in
place, but it can be useful to interleave them sometimes. Imagine this:
$c - $a
/
$e
\
$d - $b
Here $e depends on $c, etc. The optimal order may interleave the two
arms here, e.g. $a, $b, $d, $c, $e. That is because the dependencies define
a partial order, and so the arms here are actually independent.
Sorting by toplogical depth first might help in some cases, but also is not
optimal in general, as we may want to mix toplogical depths:
$a, $c, $b, $d, $e does so, and it may be the best ordering.
This PR implements a natural greedy algorithm that picks the global with
the highest use count at each step, out of the set of possible globals, which
is the set of globals that have no unresolved dependencies. So we start by
picking the first global with no dependencies and add at at the front; then
that unlocks anything that depended on it and we pick from that set, and
so forth.
This may also not be optimal, but it is easy to make it more flexible by
customizing the counts, and we consider 4 sorts here:
* Set all counts to 0. This means we only take into account dependencies,
and we break ties by the original order, so this is as close to the original
order as we can be.
* Use the actual use counts. This is the simple greedy algorithm.
* Set the count of each global to also contain the counts of its children,
so the count is the total that might be unlocked. This gives more weight
to globals that can unlock more later, so it is less greedy.
* As last, but weight children's counts lower in an exponential way, which
makes sense as they may depend on other globals too.
In practice it is simple to generate cases where 1, 2, or 3 is optimal (see
new tests), but on real-world J2CL I see that 4 (with a particular exponential
coefficient) is best, so the pass computes all 4 and picks the best. As a
result it will never worsen the size and it has a good chance of
improving.
The differences between these are small, so in theory we could pick any
of them, but given they are all modifications of a single algorithm it is
very easy to compute them all with little code complexity.
The benefits are rather small here, but this can save a few hundred
bytes on a multi-MB Java file. This comes at a tiny compile time cost, but
seems worth it for the new guarantee to never regress size.
|
| |
|
|
|
|
|
|
|
| |
--log-execution=NAME will use NAME as the module for the logger
function import, rather than infer it.
If the name is not provided (--log-execution as before this PR) then we
will try to automatically decide which to use ("env", unless we see
another module name is used, which can be the case in optimized
modules).
|
| |
|
|
|
| |
If we replace a type with another, use the original name for the new type,
and give the old a unique name (for the rare cases in which it has uses).
|
| | |
|
| |
|
|
|
| |
We had that logic right in other places, but the specific part of Vacuum that
looks at code that leads up to an unreachable did not check for infinite loops,
so it could remove them.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The SignaturePruning pass optimizes away parameters that it proves are safe to
remove. It turns out that that does not always match the definition of private
types, which is more restrictive. Specifically, if say all the types are in one big
rec group and one of them is used on an exported function then all of them are
considered public (as the rec group is). However, in closed world, it would be ok
to leave that rec group unchanged but to create a pruned version of that type
and use it, in cases where we see it is safe to remove a parameter. (See the
testcase for a concrete example.)
To put it another way, SignaturePruning already proves that a parameter is
safe to remove in all the ways that matter. Before this PR, however, the testcase
in this PR would error - so this PR is not an optimization but a bugfix, really -
because SignaturePruning would see that a parameter is safe to remove but
then TypeUpdating would see the type is public and so it would leave it alone,
leading to a broken module.
This situation is in fact not that rare, and happens on real-world Java code.
The reason we did not notice it before is that typically there are no remaining
SignaturePruning opportunities late in the process (when other closed world
optimizations have typically led to a single big rec group).
The concrete fix here is to add additionalPrivateTypes to a few more places
in TypeUpdating. We already supported that for cases where a pass knew
better than the general logic what can be modified, and this adds that
ability to the signature-rewriting logic there. Then SignaturePruning can
send in all the types it has proven are safe to modify.
* Also necessary here is to only add from additionalPrivateTypes if the type
is not already in our list (or we'd end up with duplicates in the final rec
group).
* Also move newSignatures in SignaturePruning out of the top level, which
was confusing (the pass has multiple iterations, and we want each to have
a fresh instance).
|
| |
|
|
|
| |
Remove `SExpressionParser`, `SExpressionWasmBuilder`, and `cashew::Parser`.
Simplify gen-s-parser.py. Remove the --new-wat-parser and
--deprecated-wat-parser flags.
|
| |
|
|
|
|
|
|
|
| |
(#6584)
Heap stores (struct.set) are optimized into the struct.new when they are adjacent
in a statement list.
Pushing struct.new down past irrelevant instructions increases the likelihood that
it ends up adjacent to sets.
|
| |
|
|
| |
This renames old EH tests in the form of `-eh-old.wast` to
`-eh-legacy.wast`, to be clearer in names.
|
| |
|
|
| |
If we wanted to switch types in such cases we'd need to refinalize (which is likely
worth doing, though other passes should refine globals anyhow).
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
validation (#6603)
GlobalRefining did not traverse module code, so it did not update global.gets
in other globals.
Add missing validation that actually errors on that: We did not check global.get
types.
These could be separate PRs but it would be difficult to test them separately.
|
| |
|
|
|
|
|
|
|
| |
This allows modules to contains both 32-bit and 64-bit segment.
In order to check the table/memory state when visiting segments we need
to ensure that memories/tables are visited only after their segments.
The comments in visitTable/visitMemory already assumed this but it
wasn't true in practice.
|
| | |
|
| |
|
|
|
| |
Changes to wasm-validator.cpp here are mostly for consistency between
elem and data segment validation.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
test (#6596)
This existed before #6495 but became noticeable there. We only looked at
the fallthrough values in the later part of areConsecutiveInputsEqual, but
there can be invalidation due to the non-fallthrough part:
(i32.add
(local.get $x)
(block
(local.set $x ..)
(local.get $x)
)
)
The set can cause the local.get to differ the second time. To fix this,
check if the non-fallthrough part invalidates the fallthrough (but only
on the right hand side).
Fixes #6593
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We settled on the name `WASM_EXNREF` for the new setting in Emscripten
for the name for the new EH option.
https://github.com/emscripten-core/emscripten/blob/2bc5e3156f07e603bc4f3580cf84c038ea99b2df/src/settings.js#L782-L786
"New EH" sounds vague and I'm not sure if "experimental" is really
necessary anyway, given that the potential users of this option is aware
that this is a new spec that has been adopted recently.
To make the option names consistent, this renames `--translate-to-eh`
(the option that only runs the translator) to `--translate-to-exnref`,
and `--experimental-new-eh` to `--emit-exnref` (the option that runs the
translator at the end of the whole pipeline), and renames the pass and
variable names in the code accordingly as well.
In case anyone is using the old option names (and also to make the
Chromium CI pass), this does not delete the old options.
|
| |
|
|
|
|
| |
The stringref proposal has been superseded by the imported JS strings proposal,
but the former has many more operations than the latter. To reduce complexity,
remove all operations that are part of stringref but not part of imported
strings.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The stringview types from the stringref proposal have three irregularities that
break common invariants and require pervasive special casing to handle properly:
they are supertypes of `none` but not subtypes of `any`, they cannot be the
targets of casts, and they cannot be used to construct nullable references. At
the same time, the stringref proposal has been superseded by the imported
strings proposal, which does not have these irregularities. The cost of
maintaing and improving our support for stringview types is no longer worth the
benefit of supporting them.
Simplify the code base by entirely removing the stringview types and related
instructions that do not have analogues in the imported strings proposal and do
not make sense in the absense of stringviews.
Three remaining instructions, `stringview_wtf16.get_codeunit`,
`stringview_wtf16.slice`, and `stringview_wtf16.length` take stringview operands
in the stringref proposal but cannot be removed because they lower to operations
from the imported strings proposal. These instructions are changed to take
stringref operands in Binaryen IR, and to allow a graceful upgrade path for
users of these instructions, the text and binary parsers still accept but ignore
`string.as_wtf16`, which is the instruction used to convert stringrefs to
stringviews. The binary writer emits code sequences that use scratch locals and `string.as_wtf16` to keep the output valid.
Future PRs will further align binaryen with the imported strings proposal
instead of the stringref proposal, for example by making `string` a subtype of
`extern` instead of a subtype of `any` and by removing additional instructions
that do not have analogues in the imported strings proposal.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
#6587 was incorrect: It checked generativity early in an incremental manner, but
it did not accumulate that information as we do with hashes. As a result we
could end up optimizing something with a generative child, and sadly we lacked
testing for that case.
This adds incremental generativity computation alongside hashes. It also splits
out this check from isRelevant.
Also add a test for nested effects (as opposed to generativity), but that already
worked before this PR (as we compute effects and invalidation as we go, already).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Given:
(ORIGINAL)
(in between)
(COPY)
We want to change that to
(local.tee $temp (ORIGINAL))
(in between)
(local.get $temp)
It is fine if "in between" traps: then we never reach the new local.get. This is a safer
situation than most optimizations because we are not reordering anything, only
replacing known-equivalent code.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously we checked late, and as a result might end up failing to optimize when
a sub-pattern could have worked. E.g.
(call
(A)
)
(call
(A)
)
The call cannot be optimized, but the A pattern repeats. Before this PR we'd
greedily focus on the entire call and then fail. After this PR we skip the call
before we commit to which patterns to try to optimize, so we succeed.
Add a isShallowlyGenerative helper here as we compute this step by step as
we go. Also remove a parameter to the generativity code (it did not use the
features it was passed).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously we had passes --generate-stack-ir, --optimize-stack-ir, --print-stack-ir
that could be run like any other passes. After generating StackIR it was stashed on
the function and invalidated if we modified BinaryenIR. If it wasn't invalidated then
it was used during binary writing. This PR switches things so that we optionally
generate, optimize, and print StackIR only during binary writing. It also removes
all traces of StackIR from wasm.h - after this, StackIR is a feature of binary writing
(and printing) logic only.
This is almost NFC, but there are some minor noticeable differences:
1. We no longer print has StackIR in the text format when we see it is there. It
will not be there during normal printing, as it is only present during binary writing.
(but --print-stack-ir still works as before; as mentioned above it runs during writing).
2. --generate/optimize/print-stack-ir change from being passes to being flags that
control that behavior instead. As passes, their order on the commandline mattered,
while now it does not, and they only "globally" affect things during writing.
3. The C API changes slightly, as there is no need to pass it an option "optimize" to
the StackIR APIs. Whether we optimize is handled by --optimize-stack-ir which is
set like other optimization flags on the PassOptions object, so we don't need the
old option to those C APIs.
The main benefit here is simplifying the code, so we don't need to think about
StackIR in more places than just binary writing. That may also allow future
improvements to our usage of StackIR.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
intialization (#6571)
Constants that need to be hoisted sometimes are initialized by calling
getters of other constants that need to be hoisted. These getters are
non-trivial, e.g.
(func $getConst1_<once>_@X (result (ref null $A))
(block (result (ref null $A))
(if (i32.eqz (ref.is_null (global.get $$const1@X)))
(then
(return (global.get $$const1@X))
)
)
(global.set $$const1@X (struct.new $A (i32.const 2)))
(global.get $$const1@X)
)
(func $getConst2_<once>_@X (result (ref null $A))
(block (result (ref null $A))
(if (i32.eqz (ref.is_null (global.get $$const2@X)))
(then
(return (global.get $$const2@X))
)
)
(global.set $$const2@X .... expression with (call $getConst1_<once>_@X) ....))
(global.get $$const2@X)
)
and can only be simplified after the constants they initialize are hoisted. After
the constant is hoisted the getter can be inlined and constants that depend on
it for their initialization can now be hoisted.
Before this pass, inlining would happen after the pass was run by a subsequent
run of the inliner (likely as part of -O3), requiring as many runs of this pass,
interleaved with the inliner, as the depth in the call sequence.
By having a simpler inliner run as part of the loop in this pass, the pass becomes
more effective and more independent of the call depths.
|
| |
|
|
| |
This makes the cleanup of bodies of functions that have had constants hoisted from them more effective.
|
| |
|
|
|
|
|
| |
When the input has branches to block scope, IR builder generally has to add a
wrapper block with a label name for the branch to target. To reduce the parsed
IR size, add a special case for when the wrapped expression is already an
unnamed block. In that case we can simply add the label to the existing block
instead of creating a new wrapper block.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
(#6549)
As suggested in #6434 (comment) , lower ref.cast of string views
to ref.as_non_null in binary writing. It is a simple hack that avoids the
problem of V8 not allowing them to be cast.
Add fuzzing support for the last three core string operations, after which
that problem becomes very frequent.
Also add yet another makeTrappingRefUse that was missing in that
fuzzer code.
|
| |
|
|
|
|
|
|
| |
We added string.compare late in the spec process, and forgot to add effects for it.
Unlike string.eq, it can trap.
Also use makeTrappingRefUse in recent fuzzer string generation places that I
forgot, which should reduce the amount of traps in fuzzer output.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new text parser is faster and more standards compliant than the old text
parser. Enable it by default in wasm-opt and update the tests to reflect the
slightly different results it produces. Besides following the spec, the new
parser differs from the old parser in that it:
- Does not synthesize `loop` and `try` labels unnecessarily
- Synthesizes different block names in some cases
- Parses exports in a different order
- Parses `nop`s instead of empty blocks for empty control flow arms
- Does not support parsing Poppy IR
- Produces different error messages
- Cannot parse `pop` except as the first instruction inside a `catch`
|
| |
|
|
|
|
| |
Normally a bottom type cannot reach there, as we ignore unreachable GC
operations early on. However, we can infer a bottom type later during the
flow, so we need to handle that (just not error on it, and for clarity during
debugging we also clear the contents).
|
| |
|
|
|
| |
We were reusing mutable globals in StringGathering, which meant
that we'd use a global to represent a particular string but if it was mutated
then it could contain a different string during execution.
|
