| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
| |
(#1017)
* Extends wasm-as, wasm-dis and s2wasm to consume debug locations.
* Exports source map from asm2wasm
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
* values cannot flow through an if without an else, they never return a value
* check pass tests in pass-debug mode too
* add missing finalization in binary reading
|
| |
|
|
|
|
| |
* address review feedback for #1014
|
| | |
|
| |
|
|
|
|
|
| |
* validate that types are properly finalized, when in pass-debug mode (BINARYEN_PASS_DEBUG env var): check after each pass is run that the type of each node is equal to the proper type (when finalizing it, i.e., fully recomputing the type).
* fix many fuzz bugs found by that.
* in particular, fix dce bugs with type changes not being fully updated during code removal. add a new TypeUpdater helper class that lets a pass update types efficiently, by the helper tracking deps between blocks and branches etc., and updating/propagating type changes only as necessary.
|
| | |
|
| |
|
|
|
| |
This adds a pass that converts to a CFG, runs the relooper, and re-generates wasm from that. This depends on flatten-control-flow being run before.
The main goal here is to help code generators other than asm2wasm (which already receives relooped code from fastcomp).
|
| |
|
|
|
|
|
|
| |
* parallelize istring creation, by having a thread-local set and a global set guarded by a mutex. each time a new string shows up in a thread, it will be added to that thread's set, after accessing the global set through the lock first, which means we lock at most once per new string per thread
* don't leak strings in istring store
* since we now create names in a parallel thread-safe manner, we don't need to pre-create names in RelooperJumpThreading
|
| | |
|
| |
|
|
| |
unnecessary complexity in that pass as well
|
| |
|
|
|
|
|
|
|
|
|
| |
This pass flattens out control flow in order to achieve 2 properties:
* Control flow structures (block, loop, if) and control flow operations (br, br_if, br_table, return, unreachable) may only be block children, a loop body, or an if-true or if-false. (I.e., they cannot be nested inside an i32.add, a drop, a call, an if-condition, etc.)
* Disallow block, loop, and if return values, i.e., do not use control flow to pass around values.
As a result, expressions cannot contain control flow, and overall control flow is simpler, more structured, and more "flat".
This should make things like re-relooping wasm code much easier, as they can run after the cfg is flattened
|
| |
|
|
| |
but does not properly modify the type (#1000)
|
| |
|
|
| |
the transformation, as the outside might care about that
|
| |
|
|
|
|
|
|
|
|
| |
* improve dce to handle more cases of nested unreachable code, in particular, when the child is unreachable in type but not an actual Unreachable node, e.g. if it's a br. in that case, we just need to verify that the br is not to us where we are a block or loop
* handle unreachable switch conditions in dce
* handle dce of br condition which is unreachable, and host arguments
* handle dce of block i32 etc. which is actually unreachable
|
| |
|
|
| |
wasm added if types, which mean the if can be i32 even if the arms are unreachable etc (#991)
|
| |
|
|
|
|
|
|
| |
* Add pass to instrument loads / stores
* Simplify instrumentation.
* Document.
|
| |
|
|
|
| |
Add wasm-ctor-eval, which evaluates functions at compile time - typically static constructor functions - and applies their effects into memory, saving work at startup. If we encounter something we can't evaluate at compile time in our interpreter, stop there.
This is similar to ctor_evaller.py in emscripten (which was for asm.js).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* add debugInfo option to passes, and use it to keep debug info alive through optimizations when we need it
* add fib testcase for debug info
* when preserving debug info, do not move code around call-imports, so debug info intrinsics remain stationary
* improve wasm-module-building handling of the single-threaded case: don't create workers, which is more efficient and also nicer for debugging
* process debug info in a more precise way, reordering it from being after the node (as it was a comment in JS) to before the node
* remove unreachable hack for debug info, which is no longer needed since we reorder them, and make sure to finalize blocks in which we reorder
|
| |
|
| |
Rather than storing debug info as text annotations, store explicit file and line information. This will make it easier to experiment with outputting other serializations or representations (e.g. source maps), and will allow outputting debug info for binaries as well.
|
| |
|
|
| |
* fix eq/ne of sign-ext with a constant, when the constant can never be equal to it as it has the effective sign bit but not the upper bits above it set, which the sign-ext would emit
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Move WasmType function implementations to wasm.cpp
* Move Literal methods to wasm.cpp
* Reorder wasm.cpp shared constants back to top
* Move expression functions to wasm.cpp
* Finish moving things to wasm.cpp
* Split out Literal into its own .h/.cpp. Also factor out common wasm-type module
* Remove unneeded/transitive includes from wasm.h
* Add comment to try/check methods
* Rename tryX/checkX methods to getXOrNull
* Add missing include that should fix appveyor build breakage
* More appveyor
|
| |
|
|
|
|
| |
* optimize pow(x,2) => x*x
* optimize pow(x, 0.5) => sqrt(x)
|
| |
|
|
|
|
| |
* fix a bug where compared sign-exts of different sizes were turned into zero-exts
* fix a bug where we consider an almost sign-ext as ok to change the sign of a load inside it, ignoring that the load has the extra shifting
|
| |
|
| |
Simple local common subexpression elimination. Useful mostly to reduce code size (as VMs do GVN etc.). Enabled by default in -Oz.
|
| |
|
|
| |
if would be invalid (#927)
|
| |
|
|
|
| |
* fix BINARYEN_PASS_DEBUG option
* Add isNested property to passRunner
|
| |
|
|
| |
Most module walkers use PostWalker<T, Visitor<T>>, let that pattern be
expressed as simply PostWalker<T>
|
| | |
|
| |
|
|
| |
else (#915)
|
| | |
|
| |
|
|
| |
PickLoadSigns pass
|
| | |
|
| |
|
|
| |
zero-extend
|
| | |
|
| |
|
|
| |
fix the maxBits of a signed load, which this uncovered - all the bits may be used in such a case
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
* optimize 'almost' sign extends: when we can remove one entirely, then extra shifts can be left behind. with that in place, we can then optimize 'squared' operations like shl on shl, as doing so does not break our sign extend opts
|
