| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
IRBuilder is a utility for turning arbitrary valid streams of Wasm
instructions into valid Binaryen IR. It is already used in the text
parser, so now use it in the binary parser as well. Since the IRBuilder
API for building each intruction requires only the information that the
binary and text formats include as immediates to that instruction, the
parser is now much simpler than before. In particular, it does not need
to manage a stack of instructions to figure out what the children of
each expression should be; IRBuilder handles this instead.
There are some differences between the IR constructed by IRBuilder and
the IR the binary parser constructed before this change. Most
importantly, IRBuilder generates better multivalue code because it
avoids eagerly breaking up multivalue results into individual components
that might need to be immediately reassembled into a tuple. It also
parses try-delegate more correctly, allowing the delegate to target
arbitrary labels, not just other `try`s. There are also a couple
superficial differences in the generated label and scratch local names.
As part of this change, add support for recording binary source
locations in IRBuilder.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before the PR:
$ bin/wasm-opt test/hello_world.wat --metrics
total
[exports] : 1
[funcs] : 1
[globals] : 0
[imports] : 0
[memories] : 1
[memory-data] : 0
[tables] : 0
[tags] : 0
[total] : 3
[vars] : 0
Binary : 1
LocalGet : 2
After the PR:
$ bin/wasm-opt test/hello_world.wat --metrics
Metrics
total
[exports] : 1
[funcs] : 1
...
Note the "Metrics" addition at the top. And the title can be customized:
$ bin/wasm-opt test/hello_world.wat --metrics=text
Metrics: text
total
[exports] : 1
[funcs] : 1
The custom title can be helpful when multiple invocations of metrics are used
at once, e.g. --metrics=before -O3 --metrics=after.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
We previously supported (and primarily used) a non-standard text format for
conditionals in which the condition, if-true expression, and if-false expression
were all simply s-expression children of the `if` expression. The standard text
format, however, requires the use of `then` and `else` forms to introduce the
if-true and if-false arms of the conditional. Update the legacy text parser to
require the standard format and update all tests to match. Update the printer to
print the standard format as well.
The .wast and .wat test inputs were mechanically updated with this script:
https://gist.github.com/tlively/85ae7f01f92f772241ec994c840ccbb1
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When printing Binaryen IR, we previously generated names for unnamed heap types
based on their structure. This was useful for seeing the structure of simple
types at a glance without having to separately go look up their definitions, but
it also had two problems:
1. The same name could be generated for multiple types. The generated names did
not take into account rec group structure or finality, so types that differed
only in these properties would have the same name. Also, generated type names
were limited in length, so very large types that shared only some structure
could also end up with the same names. Using the same name for multiple types
produces incorrect and unparsable output.
2. The generated names were not useful beyond the most trivial examples. Even
with length limits, names for nontrivial types were extremely long and visually
noisy, which made reading disassembled real-world code more challenging.
Fix these problems by emitting simple indexed names for unnamed heap types
instead. This regresses readability for very simple examples, but the trade off
is worth it.
This change also reduces the number of type printing systems we have by one.
Previously we had the system in Print.cpp, but we had another, more general and
extensible system in wasm-type-printing.h and wasm-type.cpp as well. Remove the
old type printing system from Print.cpp and replace it with a much smaller use
of the new system. This requires significant refactoring of Print.cpp so that
PrintExpressionContents object now holds a reference to a parent
PrintSExpression object that holds the type name state.
This diff is very large because almost every test output changed slightly. To
minimize the diff and ease review, change the type printer in wasm-type.cpp to
behave the same as the old type printer in Print.cpp except for the differences
in name generation. These changes will be reverted in much smaller PRs in the
future to generally improve how types are printed.
|
| |
|
|
|
|
|
|
|
|
| |
All top-level Module elements are identified and referred to by Name, but for
historical reasons element and data segments were referred to by index instead.
Fix this inconsistency by using Names to refer to segments from expressions that
use them. Also parse and print segment names like we do for other elements.
The C API is partially converted to use names instead of indices, but there are
still many functions that refer to data segments by index. Finishing the
conversion can be done in the future once it becomes necessary.
|
| |
|
|
|
|
|
| |
Previously when we parsed `string.const` payloads in the text format we were
using the text strings directly instead of un-escaping them. Fix that parsing,
and while we're editing the code, also add support for the `\r` escape allowed
by the spec. Remove a spurious nested anonymous namespace and spurious `static`s
in Print.cpp as well.
|
| |
|
|
|
|
|
| |
This PR removes the single memory restriction in IR, adding support for a single module to reference multiple memories. To support this change, a new memory name field was added to 13 memory instructions in order to identify the memory for the instruction.
It is a goal of this PR to maintain backwards compatibility with existing text and binary wasm modules, so memory indexes remain optional for memory instructions. Similarly, the JS API makes assumptions about which memory is intended when only one memory is present in the module. Another goal of this PR is that existing tests behavior be unaffected. That said, tests must now explicitly define a memory before invoking memory instructions or exporting a memory, and memory names are now printed for each memory instruction in the text format.
There remain quite a few places where a hardcoded reference to the first memory persist (memory flattening, for example, will return early if more than one memory is present in the module). Many of these call-sites, particularly within passes, will require us to rethink how the optimization works in a multi-memories world. Other call-sites may necessitate more invasive code restructuring to fully convert away from relying on a globally available, single memory pointer.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Updating wasm.h/cpp for DataSegments
* Updating wasm-binary.h/cpp for DataSegments
* Removed link from Memory to DataSegments and updated module-utils, Metrics and wasm-traversal
* checking isPassive when copying data segments to know whether to construct the data segment with an offset or not
* Removing memory member var from DataSegment class as there is only one memory rn. Updated wasm-validator.cpp
* Updated wasm-interpreter
* First look at updating Passes
* Updated wasm-s-parser
* Updated files in src/ir
* Updating tools files
* Last pass on src files before building
* added visitDataSegment
* Fixing build errors
* Data segments need a name
* fixing var name
* ran clang-format
* Ensuring a name on DataSegment
* Ensuring more datasegments have names
* Adding explicit name support
* Fix fuzzing name
* Outputting data name in wasm binary only if explicit
* Checking temp dataSegments vector to validateBinary because it's the one with the segments before we processNames
* Pass on when data segment names are explicitly set
* Ran auto_update_tests.py and check.py, success all around
* Removed an errant semi-colon and corrected a counter. Everything still passes
* Linting
* Fixing processing memory names after parsed from binary
* Updating the test from the last fix
* Correcting error comment
* Impl kripken@ comments
* Impl tlively@ comments
* Updated tests that remove data print when == 0
* Ran clang format
* Impl tlively@ comments
* Ran clang-format
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this we had special logic for various call types. This replaces all
that with a single general code path, which unifies everything except
for control flow constructs (which remain as before, handled in a
special way for each of them).
The algorithm is simple and direct, basically it goes through the
children and when it finds a block, it sees if it can move the block's
contents outside of the parent. While doing so it takes into account
effects and so forth.
To make this easy, a random-access API is added to ChildIterator.
Diff without whitespace makes the existing test updates a lot simpler.
Note that this is not NFC as the old algorithm had some quirks like
not taking into account effects when there were more than 2 children;
the new code is uniform in how it handles things.
This ends up removing 19% of all blocks in j2wasm, which reduces
1% of total code size.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The old algorithm can be summarized as: In each basic block, start at the beginning.
Each pair of live locals there might interfere with each other, as they might arrive from
different entry blocks with different values. Afterwards, go through the block and find
overlapping live ranges, and mark interferences there as well.
This is non-linear because at the start of the block we do a double-loop over all
pairs of live locals, which in general can be O(N^2) (N - number of locals). It also
has the downside of ignoring copies: if two locals have overlapping live ranges but
they must have identical values on those ranges, they do not actually interfere,
for example
x = 10;
y = x;
.. // live ranges overlap here
foo(x, y); // live ranges end here.
We can ignore this overlap since the copy shows they are identical there, but the
pass did not take this into account. To some extent other passes can remove such
copies (SimplifyLocals, MergeLocals, RedundantSetElimination), but in general
this was a weak spot for the optimizer.
I realized there is a solution to both these problems: In Wasm, given that we have
a default value for all locals, if a local is live at the start of a block then it must be
live at the end of all the blocks reaching it. That is so because the liveness will
extend backwards all the way to some set of the local, possibly all the way to
the zero-initialization at the start of the function, and it extends that way through
all predecessor blocks. A consequence of this is that there are no interferences
between locals that only occur during a merge: The live ranges include the
predecessor blocks, and theirs, and so forth, until we reach a block where one
of the locals is assigned a value different than the other. That is a necessary and
sufficient condition for intererence, and therefore when processing a block we
only need to look at its contents, and can ignore the merging of control flow,
which allows us to be linear.
More details on this and on the new algorithm in comments in the source, but
the basic idea is that it simply goes through each block in a linear way, finding
which values are assigned to each local (using a numbering of unique values),
and noting which are live at each time. If two locals are live and one is assigned
a value that is not the same as the value in the other, mark them as interfering.
This is of substantial benefit to j2wasm output, I believe because it is common
there to find local subexpression elimination opportunities after inlining, and
each time we find one we add a local. If we inline different functions into the
same target, we may end up with copied locals for each of them. (This was
not noticed in the past because it is very rare on LLVM output, which has
already had inlining and GVN etc. done.)
There is a small benefit to LLVM output as well, though just a few
percent at best. However, it is enough to be noticeable on some of
the code size tests.
This is also faster than the previous pass. It's normally not noticeable
as this pass is not one of the slowest anyhow, but I found some real-world
codebases where the pass becomes 50% faster. I have not found any
case where it is slower than the old algorithm.
Fuzzed over several days to be sure this is correct, and also verified
on the emscripten test suite.
|
| |
|
|
|
| |
Code in the If condition can be moved out to before the if.
Existing test updates are 99% whitespace.
|
| |
|
|
|
|
|
|
|
|
|
| |
We recently decided to change 'event' to 'tag', and to 'event section'
to 'tag section', out of the rationale that the section contains a
generalized tag that references a type, which may be used for something
other than exceptions, and the name 'event' can be confusing in the web
context.
See
- https://github.com/WebAssembly/exception-handling/issues/159#issuecomment-857910130
- https://github.com/WebAssembly/exception-handling/pull/161
|
| |
|
|
|
|
|
|
|
| |
As found in #3682, the current implementation of type ordering is not correct,
and although the immediate issue would be easy to fix, I don't think the current
intended comparison algorithm is correct in the first place. Rather than try to
switch to using a correct algorithm (which I am not sure I know how to
implement, although I have an idea) this PR removes Type ordering entirely. In
places that used Type ordering with std::set or std::map because they require
deterministic iteration order, this PR uses InsertOrdered{Set,Map} instead.
|
| |
|
|
|
|
|
|
|
|
|
| |
We used to print active element segments right after corresponding
tables, and passive segments came after those. We didn't print internal
segment names, and empty segments weren't being printed at all. This
meant that there was no way for instructions to refer to those table
segments after round tripping.
This will fix those issues by printing segments in the order they were
defined, including segment names when necessary and not omitting
empty segments anymore.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This implements emscripten-core/emscripten#13744
Inlining functions with a single use allows us to remove the function afterward.
That looks highly beneficial, shrinking every single benchmark in emscripten's
benchmark suite, by an average of 2% on the macrobenchmarks and 3.5% on
all of them. Speed also improves, although mostly on the microbenchmarks so
that might be less realistic.
There may be a slight downside to startup time due to emitting larger functions,
but given the baseline compilers in VMs these days it seems worth it, as the
delay would be just to get to the upper tier. On the benchmark suite the risk
seems low.
See more details in the PR above.
|
| |
|
|
|
|
| |
This PR adds support for `ref.null t` as a valid element segment
item. The abbreviated format of `(elem ... func $f $g...)` is kept in
both printing and binary emitting if all items are `ref.func`s. Public
APIs aren't updated in this PR.
|
| |
|
| |
Adds support for modules with multiple tables. Adds a field for the table name to `CallIndirect` and updates the C/JS APIs accordingly.
|
| |
|
|
|
|
|
|
|
|
| |
This used to return a simple name like "if" for an If, but it is redundant with
our proper printing logic. This PR turns it into a trivial helper that just prints
out the name of the class, so it now prints "If" with a capital. That is useful
for some logging, like in Metrics I think it is clearer than it was earlier (since
we are actually counting the classes, and our old emitting of text-format-like
names are just confusing, as we emitted "binary" there which is not valid).
Also replace some usages of that method with proper printing.
|
| |
|
|
|
|
|
| |
Using addition in more places is better for gzip, and helps simplify the
optimizer as well.
Add a FinalOptimizer phase to do optimizations like our signed LEB tweaks, to
reduce binary size in the rare case when we do want a subtraction.
|
| |
|
|
|
| |
Selectify turns an if-else into a select where possible. Previously we abandoned
hope if any part of the if had a side effect. But it's fine for the condition to have a
side effect, so long as moving it to the end doesn't invalidate the arms.
|
| |
|
|
|
|
|
|
|
| |
The s-parser was assigning numbers names per-type where as
the binaryn reader was using the global import count as the
number to append.
This change switches to use per-element count which I think
it preferable as it increases the stability of the auto-generated
names. e.g. memory is now always named `$mimport0`.
|
| |
|
| |
Implements the parts of the Extended Name Section Proposal that are trivially applicable to Binaryen, in particular table, memory and global names. Does not yet implement label, type, elem and data names.
|
| |
|
|
|
|
|
|
| |
`BinaryIndexes` was only used in two places (Print.cpp and
wasm-binary.h), so it didn't seem to be a great fit for
module-utils.h. This change moves it to wasm-binary.h and removes its
usage in Print.cpp. This means that function indexes are no longer
printed, but those were of limited utility and were the source of
annoying noise when updating tests, anyway.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Function signatures were previously redundantly stored on Function
objects as well as on FunctionType objects. These two signature
representations had to always be kept in sync, which was error-prone
and needlessly complex. This PR takes advantage of the new ability of
Type to represent multiple value types by consolidating function
signatures as a pair of Types (params and results) stored on the
Function object.
Since there are no longer module-global named function types,
significant changes had to be made to the printing and emitting of
function types, as well as their parsing and manipulation in various
passes.
The C and JS APIs and their tests also had to be updated to remove
named function types.
|
| |
|
|
|
|
|
|
|
| |
This is the start of a larger refactoring to remove FunctionType entirely and
store types and signatures directly on the entities that use them. This PR
updates BrOnExn and Events to remove their use of FunctionType and makes the
BinaryWriter traverse the module and collect types rather than using the global
FunctionType list. While we are collecting types, we also sort them by frequency
as an optimization. Remaining uses of FunctionType in Function, CallIndirect,
and parsing will be removed in a future PR.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds support for the event and the event section, as specified in
https://github.com/WebAssembly/exception-handling/blob/master/proposals/Exceptions.md#changes-to-the-binary-model.
Wasm events are features that suspend the current execution and transfer
the control flow to a corresponding handler. Currently the only
supported event kind is exceptions.
For events, this includes support for
- Binary file reading/writing
- Wast file reading/writing
- Binaryen.js API
- Fuzzer
- Validation
- Metadce
- Passes: metrics, minify-imports-and-exports,
remove-unused-module-elements
|
| | |
|
| |
|
|
| |
Also, always output high level metrics even when zero.
|
| |
|
|
|
|
| |
Automated renaming according to
https://github.com/WebAssembly/spec/issues/884#issuecomment-426433329.
|
| |
|
| |
If an if is enclosed in a block which is only used to exit one arm, move it into that arm, so it can be better optimized. Similar to what we did for loops in #1736.
|
| |
|
|
|
| |
That is the correct order in the text format, wabt errors otherwise.
See AssemblyScript/assemblyscript#310
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes #1649
This moves us to a single object for functions, which can be imported or nor, and likewise for globals (as a result, GetGlobals do not need to check if the global is imported or not, etc.). All imported things now inherit from Importable, which has the module and base of the import, and if they are set then it is an import.
For convenient iteration, there are a few helpers like
ModuleUtils::iterDefinedGlobals(wasm, [&](Global* global) {
.. use global ..
});
as often iteration only cares about imported or defined (non-imported) things.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a pass to remove unnecessary call arguments in an LTO-like manner, that is:
* If a parameter is not actually used in a function, we don't need to send anything, and can remove it from the function's declaration. Concretely,
(func $a (param $x i32)
..no uses of $x..
)
(func $b
(call $a (..))
)
=>
(func $a
..no uses of $x..
)
(func $b
(call $a)
)
And
* If a parameter is only ever sent the same constant value, we can just set that constant value in the function (which then means that the values sent from the outside are no longer used, as in the previous point). Concretely,
(func $a (param $x i32)
..may use $x..
)
(func $b
(call $a (i32.const 1))
(call $a (i32.const 1))
)
=>
(func $a
(local $x i32)
(set_local $x (i32.const 1)
..may use $x..
)
(func $b
(call $a)
(call $a)
)
How much this helps depends on the codebase obviously, but sometimes it is pretty useful. For example, it shrinks 0.72% on Unity and 0.37% on Mono. Note that those numbers include not just the optimization itself, but the other optimizations it then enables - in particular the second point from earlier leads to inlining a constant value, which often allows constant propagation, and also removing parameters may enable more duplicate function elimination, etc. - which explains how this can shrink Unity by almost 1%.
Implementation is pretty straightforward, but there is some work to make the heavy part of the pass parallel, and a bunch of corner cases to avoid (can't change a function that is exported or in the table, etc.). Like the Inlining pass, there is both a standard and an "optimizing" version of this pass - the latter also optimizes the functions it changes, as like Inlining, it's useful to not need to re-run all function optimizations on the whole module.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a new IR, "Stack IR". This represents wasm at a very low level, as a simple stream of instructions, basically the same as wasm's binary format. This is unlike Binaryen IR which is structured and in a tree format.
This gives some small wins on binary sizes, less than 1% in most cases, usually 0.25-0.50% or so. That's not much by itself, but looking forward this prepares us for multi-value, which we really need an IR like this to be able to optimize well. Also, it's possible there is more we can do already - currently there are just a few stack IR optimizations implemented,
DCE
local2stack - check if a set_local/get_local pair can be removed, which keeps the set's value on the stack, which if the stars align it can be popped instead of the get.
Block removal - remove any blocks with no branches, as they are valid in wasm binary format.
Implementation-wise, the IR is defined in wasm-stack.h. A new StackInst is defined, representing a single instruction. Most are simple reflections of Binaryen IR (an add, a load, etc.), and just pointers to them. Control flow constructs are expanded into multiple instructions, like a block turns into a block begin and end, and we may also emit extra unreachables to handle the fact Binaryen IR has unreachable blocks/ifs/loops but wasm does not. Overall, all the Binaryen IR differences with wasm vanish on the way to stack IR.
Where this IR lives: Each Function now has a unique_ptr to stack IR, that is, a function may have stack IR alongside the main IR. If the stack IR is present, we write it out during binary writing; if not, we do the same binaryen IR => wasm binary process as before (this PR should not affect speed there). This design lets us use normal Passes on stack IR, in particular this PR defines 3 passes:
Generate stack IR
Optimize stack IR (might be worth splitting out into separate passes eventually)
Print stack IR for debugging purposes
Having these as normal passes is convenient as then they can run in parallel across functions and all the other conveniences of our current Pass system. However, a downside of keeping the second IR as an option on Functions, and using normal Passes to operate on it, means that we may get out of sync: if you generate stack IR, then modify binaryen IR, then the stack IR may no longer be valid (for example, maybe you removed locals or modified instructions in place etc.). To avoid that, Passes now define if they modify Binaryen IR or not; if they do, we throw away the stack IR.
Miscellaneous notes:
Just writing Stack IR, then writing to binary - no optimizations - is 20% slower than going directly to binary, which is one reason why we still support direct writing. This does lead to some "fun" C++ template code to make that convenient: there is a single StackWriter class, templated over the "mode", which is either Binaryen2Binary (direct writing), Binaryen2Stack, or Stack2Binary. This avoids a lot of boilerplate as the 3 modes share a lot of code in overlapping ways.
Stack IR does not support source maps / debug info. We just don't use that IR if debug info is present.
A tiny text format comment (if emitting non-minified text) indicates stack IR is present, if it is ((; has Stack IR ;)). This may help with debugging, just in case people forget. There is also a pass to print out the stack IR for debug purposes, as mentioned above.
The sieve binaryen.js test was actually not validating all along - these new opts broke it in a more noticeable manner. Fixed.
Added extra checks in pass-debug mode, to verify that if stack IR should have been thrown out, it was. This should help avoid any confusion with the IR being invalid.
Added a comment about the possible future of stack IR as the main IR, depending on optimization results, following some discussion earlier today.
|
| |
|
|
| |
helpful in both positions on general code (#1581)
|
| |
|
|
|
| |
Add a version of simplify-locals which does not create nesting. This keeps the IR flat (in the sense of --flatten).
Also refactor simpify-locals to be a template, so the various modes are all template parameters.
|
|
|
The option keeps running the passes (that we were told to run) in cycles until we converge in terms of the binary size, that is, keep optimizing until we can't shrink any more.
Also fix a --metrics bug this uncovered: we can't expect the Metrics object to still be around if running passes later in another PassRunner.
|
