This patch updates internal data structures used by V8 to support
multiple indirect function tables (WebAssembly/design#682). But, since
this feature is post-MVP, the functionality is not directly exposed and
parsing/generation of WebAssembly is left unchanged. Nevertheless, it
is being used in an experiment to implement fine-grained control flow
integrity based on C/C++ types.
BUG=
Review-Url: https://codereview.chromium.org/2174123002
Cr-Commit-Position: refs/heads/master@{#38110}
- Add Simd128 type to Wasm AST types
- Add a pass that converts SIMD machine ops to runtime calls
- Sample opcodes Int32x4Splat, Int32x4ExtractLane and test
- Separate out generic SIMD Machine ops as these cannot be
handled by runtime functions just yet.
LOG=N
BUG=v8:4124
R=bradnelson@chromium.org, bbudge@chromium.org, titzer@chromium.org
Review-Url: https://codereview.chromium.org/1991143002
Cr-Commit-Position: refs/heads/master@{#37789}
To correctly support instantiating a compiled module multiple times, we clone the
compiled module each time we create an instance, since some of the data is specific
to the instance - e.g. export code, wasm functions, indirect table.
BUG=v8:5072
Review-Url: https://codereview.chromium.org/2134593002
Cr-Commit-Position: refs/heads/master@{#37692}
In the current implementation of wasm an unrepresentable input of the
float32-to-int32 conversion is detected by first truncating the input, then
converting the truncated input to int32 and back to float32, and then checking
whether the result is the same as the truncated input.
This input check does not work on arm and arm64 for an input of (INT32_MAX + 1)
because on these platforms the float32-to-int32 conversion results in INT32_MAX
if the input is greater than INT32_MAX. When INT32_MAX is converted back to
float32, then the result is (INT32_MAX + 1) again because INT32_MAX cannot be
represented precisely as float32, and rounding-to-nearest results in (INT32_MAX
+ 1). Since (INT32_MAX + 1) equals the truncated input value, the input appears
to be representable.
With the changes in this CL, the result of the float32-to-int32 conversion is
incremented by 1 if the original result was INT32_MAX. Thereby the detection of
unrepresenable inputs in wasm works. Note that since INT32_MAX cannot be
represented precisely in float32, it can also never be a valid result of the
float32-to-int32 conversion.
@v8-mips-ports, can you do a similar implementation for mips?
R=titzer@chromium.org, Rodolph.Perfetta@arm.com
Review-Url: https://codereview.chromium.org/2105313002
Cr-Commit-Position: refs/heads/master@{#37448}
Reland of https://codereview.chromium.org/2034093002 (reverted by
https://codereview.chromium.org/2080153002).
Original commit message:
Implement WASM support on big-endian platforms. WASM has
an implicit requirement that it is running on little-endian
machine. We achieve WASM support on BE by keeping data
in memory in little-endian order, and changing data
endianness before storing to memory and after loading from
memory.
BUG=
Review-Url: https://codereview.chromium.org/2083523002
Cr-Commit-Position: refs/heads/master@{#37373}
This reverts commit 1eb1dfabe4.
The original compilation separation change avoided associating a heap
for the wasm instance if memory was not provided, nor needed. The
grow memory CL assumed the old behavior, where a memory buffer was
always present, but may have had a zero size.
The 2CLS landed shortly after one another. We decided to treat the
grow memory as the race condition winner, so this CL here re-lands
compilation separation, plus adjusts grow memory to deal with
the undefined mem buffer.
BUG=
Review-Url: https://codereview.chromium.org/2102193003
Cr-Commit-Position: refs/heads/master@{#37352}
Support for serializing/deserializing the compiled wasm module.
We want to reuse the javascript snapshotting mechanics, at least in the
short term, when we still use the JS heap for the compiled wasm code.
Given that a module may be compiled in one v8 instance and then
instantiated later, in a different instance, whatever information we need
at instantiation time must also be serializable.
We currently hold on to the un-decoded wasm bytes, for enabling
debugging scenarios. This imposes a ~20% penalty on the memory
requirements of the wasm compiled code. We do not need this data
otherwise, for runtime, and it is sensible to consider eventually loading it
on demand. Therefore, I intentionally avoided relying on it and re-
decoding the wasm module data, and instead saved the information
necessary to support instantiation.
Given how whatever we need to persist must be serializable, the CL
uses a structure made out of serializable objects (fixed arrays mostly)
for storing this information. I preferred going this route rather than
adding more wasm-specific support to the serializer, given that we want
to eventually move off the JS heap, and therefore the serializer.
Additionally, it turns out this extra information is relatively not complex:
minimal structure, little nesting depth, mostly simple data like numbers
or byte blobs, or opaque data like compiled functions.
This CL also moves export compilation ahead of instantiation time.
This change added a helper getter to FixedArray, to make typed retrieval
of elements easier.
BUG=
Review-Url: https://codereview.chromium.org/2094563002
Cr-Commit-Position: refs/heads/master@{#37348}
Remove second wasm module compilation and instantiation path that
we had in CompileAndRunWasmModule and reuse the same path used
by user code.
BUG=
Review-Url: https://codereview.chromium.org/2091533002
Cr-Commit-Position: refs/heads/master@{#37203}
Yanking out in an effort to reduce dependencies. We probably want to
separate codegen into instance-specific and module-generic purpose -
eventually.
BUG=
Review-Url: https://codereview.chromium.org/2085863003
Cr-Commit-Position: refs/heads/master@{#37126}
Port 481502dad9
Float32SubMinusZero and Float64SubMinusZero tests are failing because MIPS does not preserve NaN payload according to Wasm spec. Implemented macro-assembler methods that check for NaN operands, and return the qNaN value with preserved payload and sign bits.
TEST=cctest/test-run-wasm/Run_WasmFloat32SubMinusZero, cctest/test-run-wasm/Run_WasmFloat64SubMinusZero
BUG=
patch from issue 2019693002 at patchset 140001 (http://crrev.com/2019693002#ps140001)
R=ahaas@chromium.org
Review-Url: https://codereview.chromium.org/2066483008
Cr-Commit-Position: refs/heads/master@{#37105}
Reason for revert:
[Sheriff] Speculative revert for crashes on chrubuntu chromebooks:
https://build.chromium.org/p/client.v8.ports/builders/V8%20Arm/builds/320
Original issue's description:
> Implement WASM big-endian support.
>
> Implement WASM support on big-endian platforms. WASM has
> an implicit requirement that it is running on little-endian
> machine. We achieve WASM support on BE by keeping data
> in memory in little-endian order, and changing data
> endianness before storing to memory and after loading from
> memory.
>
> BUG=
>
> Committed: https://crrev.com/d3f3f6c8186b2a53f0c539f7bba0c3708c4d83f9
> Cr-Commit-Position: refs/heads/master@{#37065}
TBR=titzer@chromium.org,akos.palfi@imgtec.com,balazs.kilvady@imgtec.com,jyan@ca.ibm.com,ivica.bogosavljevic@imgtec.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=
Review-Url: https://codereview.chromium.org/2080153002
Cr-Commit-Position: refs/heads/master@{#37091}
Compilation of wasm functions happens before instantiation. Imports are linked afterwards, at instantiation time. Globals and memory are also
allocated and then tied in via relocation at instantiation time.
This paves the way for implementing Wasm.compile, a prerequisite to
offering the compiled code serialization feature.
Currently, the WasmModule::Compile method just returns a fixed array
containing the code objects. More appropriate modeling of the compiled module to come.
Opportunistically centralized the logic on how to update memory
references, size, and globals, since that logic is the exact same on each
architecture, except for the actual storing of values back in the
instruction stream.
BUG=v8:5072
Review-Url: https://codereview.chromium.org/2056633002
Cr-Commit-Position: refs/heads/master@{#37086}
Implement WASM support on big-endian platforms. WASM has
an implicit requirement that it is running on little-endian
machine. We achieve WASM support on BE by keeping data
in memory in little-endian order, and changing data
endianness before storing to memory and after loading from
memory.
BUG=
Review-Url: https://codereview.chromium.org/2034093002
Cr-Commit-Position: refs/heads/master@{#37065}
We report the byte offset as column number, but devtools assumes them
to be 1-based and subtracts one unconditionally before further
processing it. It's a bit unfortunate, but because of that we have to
just add 1 to the reported column number on the public StackTrace API.
R=ahaas@chromium.org, titzer@chromium.org
BUG=chromium:613110
Review-Url: https://codereview.chromium.org/2071563002
Cr-Commit-Position: refs/heads/master@{#37036}
The last 4 test cases in test/cctest/wasm/test-run-wasm-asmjs.cc added by the CL 36911 (https://codereview.chromium.org/2061583002) use float_t and double_t type for WasmRunner.
For examples: At line 249: WasmRunner<float_t> r(&module, MachineType::Uint32());
But float_t and double_t depends on FLT_EVAL_METHOD macro of compiler. FLT_EVAL_METHOD is variant on different platform, if the FLT_EVAL_METHOD is 2, both float_t and double_t will be long
double and gcc or clang will met error when compiling WasmRunner<long double> r(&module,MachineType::Uint32());
For more details, please refer:
float_t: http://www.cplusplus.com/reference/cmath/float_t/
FLT_EVAL_METHOD: https://en.wikipedia.org/wiki/C99 check the IEEE 754 floating point support section directly.
This CL used float and double to replace float_t and double_t to avoid this issue.
BUG=
Review-Url: https://codereview.chromium.org/2066703003
Cr-Commit-Position: refs/heads/master@{#36982}
Support for relocatable globals, to facilitate compilation before
instantiation.
BUG=v8:5072
Review-Url: https://codereview.chromium.org/2062003002
Cr-Commit-Position: refs/heads/master@{#36978}
MemoryOperands, on IA32. This needed to be fixed so that we can
compile wasm code before creating instances, since the compiled code
needs to be patched up for memory and globals references.
This surfaces in asm-to-wasm scenarios.
Added testing (rather, enhanced existing tests).
Note patch#1 where we fail on ia32, and patch#2 with the fix.
BUG=v8:5072
Review-Url: https://codereview.chromium.org/2061583002
Cr-Commit-Position: refs/heads/master@{#36911}
This interpreter directly decodes and executes WASM binary code for
the purpose of supporting low-level debugging. It is not currently
integrated into the main WASM implementation.
R=ahaas@chromium.org,clemensh@chromium.org,rossberg@chromium.org,binji@chromium.org
BUG=
Review-Url: https://codereview.chromium.org/1972153002
Cr-Commit-Position: refs/heads/master@{#36497}
Moved globals offsets calculation to the wasm module decoder, since
this is a property of the module, not of each instance.
Qualified as "const" references to WasmModule outside of the decoder
and some test situations.
BUG=
Review-Url: https://codereview.chromium.org/2005933003
Cr-Commit-Position: refs/heads/master@{#36484}
Empty function names are allowed and are output as such, unnamed
functions or functions with no valid UTF-8 name are output as
"<WASM UNNAMED>", while the CallSite object returns null as the
function name.
R=titzer@chromium.org, yangguo@chromium.org
Review-Url: https://codereview.chromium.org/1970503004
Cr-Commit-Position: refs/heads/master@{#36348}
This is a first step to removing the support for the OldFunctions
section altogether, which will greatly simplify the encoder and remove
the need to do local variable remapping in asm->wasm.
R=bradnelson@chromium.org,aseemgarg@chromium.org
BUG=
Review-Url: https://codereview.chromium.org/1974933002
Cr-Commit-Position: refs/heads/master@{#36228}
