With just five cache registers, Liftoff can run out of memory on a
64bit shift. This CL solves this by using a parallel register move and
pinning less registers.
R=ahaas@chromium.org
Bug: chromium:894307
Change-Id: I91ed0fee00ceb452841e5d1bb10905be6702dcce
Reviewed-on: https://chromium-review.googlesource.com/c/1337580
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#57552}
The flag is on by default, so we don't need to specify it. More
importantly, the tests are expected to work for any value of that flag.
So don't force the flag but use whatever the test variant chooses.
Note that in streaming-compile.js, the flag was accidentally specified
as '-async-compilation'. I also removed that one.
R=ahaas@chromium.org
Change-Id: Ifad31160d266dda38cdd9dd1d73dad69bd2c2f2c
Reviewed-on: https://chromium-review.googlesource.com/c/1325961
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#57406}
For memory limit checks, we should use the minimum of the
--wasm-max-mem-pages flag and kV8MaxWasmMemoryPages. The former is a
limit set by the user, the latter is the maximum we can handle
internally.
R=titzer@chromium.org
Bug: chromium:898677
Change-Id: I3c549f4e90dd016b5d07475d9353f30134f76dcc
Reviewed-on: https://chromium-review.googlesource.com/c/1305274
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#57127}
The "grow_memory" opcode was renamed to "memory.grow", and the spec
repo was updated to use kExprMemoryGrow internally instead of
kExprGrowMemory (https://github.com/WebAssembly/spec/pull/720).
This CL does the same change for v8.
Drive-by: Rename "current_size" to "memory.size", and a minor cleanup
in wasm-graph-builder.js to bring it in line with the version in the
js-api tests in the spec repo.
R=titzer@chromium.org
Change-Id: If525dba898b2c248890a616d3392c22b45f698ef
Reviewed-on: https://chromium-review.googlesource.com/c/1302057
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#57089}
On ia32, we were pinning too many registers, resulting in no unpinned
byte registers left (we only have three byte registers since {ebx}
is reserved for the root register).
It turns out that on most paths, we don't actually need to pin any
registers, since {Store} is often the last call for an operation (like
any store or set_global). If registers need to be pinned, only pass
those that must be kept alive across the {Store}. This allows to
compute a more narrow set of pinned registers on demand inside {Store}.
Plus minor drive-by changes.
R=titzer@chromium.org
Bug: chromium:894374, chromium:894307, v8:6600
Change-Id: Ic4d7131784c193dc7a2abf0e504d9973f6d5c5f1
Reviewed-on: https://chromium-review.googlesource.com/c/1275819
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#56587}
Use naming similar to the spec: "table" instead of "function table",
"element segment" instead of "function table init".
Change-Id: Ib1b6cdfa566f8bd00017ccedf9440084204f10ff
Reviewed-on: https://chromium-review.googlesource.com/c/1273612
Commit-Queue: Ben Smith <binji@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#56545}
Previously we only supported strings and not filenames. This
changes the default to filename and adds a new `type: string` which can
be passed `options` to allow for strings to be passed in test code.
See: https://developer.mozilla.org/en-US/docs/Web/API/Worker/Worker
Bug: v8:8020
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: Ie8818885c5c5c071b6614852322cb45aeb01a647
Reviewed-on: https://chromium-review.googlesource.com/1185980
Commit-Queue: Sam Clegg <sbc@chromium.org>
Reviewed-by: Ben Smith <binji@chromium.org>
Cr-Commit-Position: refs/heads/master@{#56056}
This verifies that asm.js over the internal parameter count limit
does not crash. The internal limit is 1000 parameters, and the test
was using >3000 parameters. Reduce this down to 1005, and also
introduce a test which does not dynamically construct the string
and eval it, because the construction of this string takes time.
Mark the old test as slow in debug mode.
R=machenbach@chromium.org
BUG=v8:8165
Change-Id: Ib6ef5e1e58d3f37a71720fb59afa19464e7f2ff7
Reviewed-on: https://chromium-review.googlesource.com/1224057
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55860}
This fixes exception creation (by the WebAssembly throw operation) so
that it is not observable by JavaScript. Internal properties are now
stored with symbol names instead of string names, which also prevents
them from being accessed or monkey-patched directly by JavaScript.
R=clemensh@chromium.org
TEST=mjsunit/regress/wasm/regress-8094
BUG=v8:8094
Change-Id: I33cb27f4373114cd4db28d9aef23560093e55242
Reviewed-on: https://chromium-review.googlesource.com/1203951
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55602}
This fixes a crash with a predicate used during stack unwinding of
WebAssembly frames during exception handling. The predicate caused an
observable side-effect in JavaScript during unwinding, code that is
inherently unhandlified and is not allowed to be observable.
The fix actually just removes the entire predicate. This is because the
updated proposal causes all JavaScript exceptions to participate in
WebAssembly exception handling, allowing modelling of "finally" language
constructs to perform cleanup independent of the embedders exception
details.
R=ahaas@chromium.org
TEST=mjsunit/regress/wasm/regress-8095
BUG=v8:8095
Change-Id: Ic03bc45e7b7f4562a431ccf910ee9ddcf558aa48
Reviewed-on: https://chromium-review.googlesource.com/1193445
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55457}
This makes sure the aforementioned predicate is independent of the
current context (aka. Realm) and only uses the instance type of the
given object to determine whether it is a WebAssembly module object.
R=titzer@chromium.org
TEST=mjsunit/regress/wasm/regress-8059
BUG=v8:8059
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Icc8e400f8412483f2a3883ca65c58b7ef938ef23
Reviewed-on: https://chromium-review.googlesource.com/1180886
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55205}
Enforce both engine limitations and spec (http://asmjs.org/spec/latest/)
limitations on the size of asm.js heaps.
R=clemensh@chromium.org
CC=mstarzinger@chromium.org
Bug: chromium:873600
Change-Id: I104c23bbd0a9a7c494f97f8f9e83ac5a37496dfd
Reviewed-on: https://chromium-review.googlesource.com/1174411
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55163}
i32 stack parameters can be loaded by Turbofan as 64-bit value, hence
they would not be zero extended. If this loaded value is then passed to
Liftoff (which assumes zero-extended i32 values), we could use it for
memory accesses, which would be out of bounds.
R=mstarzinger@chromium.org
Bug: chromium:864509, v8:6600
Change-Id: I0f45a269b1fb1c2befc2e6bc660c559a88323767
Reviewed-on: https://chromium-review.googlesource.com/1140168
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54500}
This removes an occurrence where the "%Foo" native syntax appears as part
of a string. Such strings are picked up by the fuzzer and recombined in
unsupported ways, producing false-positive crash reports. Simply avoid
having those strings in the fuzzing corpus.
R=clemensh@chromium.org
TEST=mjsunit/regress/wasm/regress-808848
BUG=chromium:844842
Change-Id: I017c1552578f0d26033e58b11353e87e27a69ebf
Reviewed-on: https://chromium-review.googlesource.com/1136300
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54432}
We need to push the sign-extended constant instead of just the lower 32
bits. Otherwise, the callee might read stale data from the stack.
Bug: chromium:854011, v8:6600
R=ahaas@chromium.orgCC=rodolph.perfetta@arm.com
Change-Id: Iafcfd6ba9532771615b41215fb4d1a2b85ce5623
Reviewed-on: https://chromium-review.googlesource.com/1124683
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54185}
This is a reland of ada648006b, fixed
for 32 bit architectures (register pairs).
Original change's description:
> [Liftoff] Fix register use count
>
> In {SetLocalFromStackSlot}, we decrement the use count of the register
> in the target slot without updating this slot, and then call
> {GetUnusedRegister}. At that point, the register use counts do not
> match the cache state, which leads to errors later on.
> This CL fixes this by marking the target slot as a stack slot after
> reducing the register use count.
>
> It also adds a Validation which helped to find that error and will
> catch similar errors earlier.
>
> R=titzer@chromium.org
>
> Bug: chromium:854050, v8:6600
> Change-Id: I74d3a5aa947ec4247d7b4557567f642bf4082316
> Reviewed-on: https://chromium-review.googlesource.com/1111958
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53976}
TBR=titzer@chromium.org
Bug: chromium:854050, v8:6600
Change-Id: Ibc8801737e9604a8490382c569b0378585625376
Reviewed-on: https://chromium-review.googlesource.com/1112238
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53981}
This reverts commit ada648006b.
Reason for revert: Failure with slow dchecks: https://ci.chromium.org/p/v8/builders/luci.v8.ci/V8%20Linux%20-%20debug/20982
Original change's description:
> [Liftoff] Fix register use count
>
> In {SetLocalFromStackSlot}, we decrement the use count of the register
> in the target slot without updating this slot, and then call
> {GetUnusedRegister}. At that point, the register use counts do not
> match the cache state, which leads to errors later on.
> This CL fixes this by marking the target slot as a stack slot after
> reducing the register use count.
>
> It also adds a Validation which helped to find that error and will
> catch similar errors earlier.
>
> R=titzer@chromium.org
>
> Bug: chromium:854050, v8:6600
> Change-Id: I74d3a5aa947ec4247d7b4557567f642bf4082316
> Reviewed-on: https://chromium-review.googlesource.com/1111958
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53976}
TBR=titzer@chromium.org,clemensh@chromium.org
Change-Id: I5b8d8d405dcd7f82ee431cba290419425b9859a1
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:854050, v8:6600
Reviewed-on: https://chromium-review.googlesource.com/1112277
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53979}
In {SetLocalFromStackSlot}, we decrement the use count of the register
in the target slot without updating this slot, and then call
{GetUnusedRegister}. At that point, the register use counts do not
match the cache state, which leads to errors later on.
This CL fixes this by marking the target slot as a stack slot after
reducing the register use count.
It also adds a Validation which helped to find that error and will
catch similar errors earlier.
R=titzer@chromium.org
Bug: chromium:854050, v8:6600
Change-Id: I74d3a5aa947ec4247d7b4557567f642bf4082316
Reviewed-on: https://chromium-review.googlesource.com/1111958
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53976}
Cleanup decoding of flags so that invalid flags for sections other than
memory are caught correctly.
Bug: chromium:853453
Change-Id: Ia347d5f7672eee93ca3f6a743f06fba629f55cb5
Reviewed-on: https://chromium-review.googlesource.com/1104976
Commit-Queue: Deepti Gandluri <gdeepti@chromium.org>
Reviewed-by: Ben Smith <binji@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53972}
This loads references to {null} values from the instance object instead
of embedding them into the generated code. It is one step towards making
the {WasmCode} objects independent of the Isolate.
Note that this also fixes an issue with the serializer/deserializer that
failed to properly serialize {null} values and accidentally collapsed
them to {undefined} values instead.
R=ahaas@chromium.org
TEST=mjsunit/regress/wasm/regress-7785
BUG=v8:7424,v8:7785
Change-Id: Ie436c2d96890e7c8c89ffe2bd4189a759254775b
Reviewed-on: https://chromium-review.googlesource.com/1070981
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53352}
At the moment, WebAssembly.instantiate(bytes) is implemented by
desugaring it to WebAssembly.compile(bytes).then(WebAssembly.instantiate).
The problem is that the {then} in this snippet is observable. With this
CL I introduce a CompilationResultResolver which allows to do the
desugaring internally and thereby make the {then} unobservable.
Unfortunately the result of WebAssembly.instantiate(bytes) is different
than the result of WebAssembly.instantiate(module). Therefore I also
introduced an InstantiationResultResolver for symmetry with
WebAssembly.compile.
R=mstarzinger@chromium.org
Bug: chromium:837417
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I2d98e03d65f2ada19041d5a9e2df5da91b24ccca
Reviewed-on: https://chromium-review.googlesource.com/1059783
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53347}
The js-to-wasm wrappers are shared across instances, so we cannot
directly call the instance-specific wasm-to-js wrappers. Instead, we
need to call via the import table.
R=titzer@chromium.org
Bug: chromium:843563
Change-Id: Ia882604f6769472fe2eb69176cbed728215ced29
Reviewed-on: https://chromium-review.googlesource.com/1064610
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53254}
The clusterfuzz issue crashes because VisitBinops expected only but 4
input operands but in the generated graph 5 input operands get created
The issue is fixed by increasing the size of the input operand buffer.
R=jarin@chromium.org
Bug: chromium:842501
Change-Id: I4bbb09a968e165e6f5a0a02d06eee97333f7aa38
Reviewed-on: https://chromium-review.googlesource.com/1056989
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53147}
This is a reland of e084eea628.
Undefined behavious was fixed in https://crrev.com/c/1051235.
Original change's description:
> Fix SourcePositionInfo for wasm
>
> In wasm we often don't have a SharedFunctionInfo associated with a
> compilation job, so we can't get a Script. Just print "unknown" in
> these cases (instead of crashing).
>
> R=titzer@chromium.org
> CC=herhut@chromium.org
>
> Bug: chromium:840757, v8:7738
> Change-Id: I850c6adfd9e07c9a0f6dd018f1a9314feb89d887
> Reviewed-on: https://chromium-review.googlesource.com/1049632
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53080}
TBR=titzer@chromium.org
Bug: chromium:840757, v8:7738
Change-Id: If04040a33766955cfed78e7c27226dd04c3f9b9f
Reviewed-on: https://chromium-review.googlesource.com/1051266
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53111}
This reverts commit e084eea628.
Reason for revert:
https://build.chromium.org/p/client.v8/builders/V8%20Linux64%20UBSanVptr/builds/3163
Original change's description:
> Fix SourcePositionInfo for wasm
>
> In wasm we often don't have a SharedFunctionInfo associated with a
> compilation job, so we can't get a Script. Just print "unknown" in
> these cases (instead of crashing).
>
> R=titzer@chromium.org
> CC=herhut@chromium.org
>
> Bug: chromium:840757, v8:7738
> Change-Id: I850c6adfd9e07c9a0f6dd018f1a9314feb89d887
> Reviewed-on: https://chromium-review.googlesource.com/1049632
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53080}
TBR=titzer@chromium.org,clemensh@chromium.org
Change-Id: Ib2020ea3f2b778df9fe50ccbe803938f2f4fd709
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:840757, v8:7738
Reviewed-on: https://chromium-review.googlesource.com/1051265
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53082}
In wasm we often don't have a SharedFunctionInfo associated with a
compilation job, so we can't get a Script. Just print "unknown" in
these cases (instead of crashing).
R=titzer@chromium.org
CC=herhut@chromium.org
Bug: chromium:840757, v8:7738
Change-Id: I850c6adfd9e07c9a0f6dd018f1a9314feb89d887
Reviewed-on: https://chromium-review.googlesource.com/1049632
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53080}
A stack overflow can be thrown by JSEntryStub, which means the
thread-in-wasm flag will not have the expected value. To accommodate
this, we now clear the flag during exceptional returns if it is set.
Bug: chromium:834624
Change-Id: I8359af79886ab98dfecc2fb39ca19118b7fa38eb
Reviewed-on: https://chromium-review.googlesource.com/1019570
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Eric Holk <eholk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52891}
When WebAssembly.instantiate or WebAssembly.instantiateStreaming is
called in JavaScript, internally we transfrom it into
WebAssembly.compile(buffer).then(WebAssembly.instantiate). However,
modifying the prototype of WebAssembly.Module can change the result of
WebAssembly.compile(buffer). With this CL we make sure that even if the
result of WebAssembly.compile is modified, there is still no type
confusion. In the long term we have to do a refactoring and remove
this internal transformation.
R=mstarzinger@chromium.org
Bug: chromium:837417
Change-Id: I376068b8b8b01b991ec450162da6a62ae7030c62
Reviewed-on: https://chromium-review.googlesource.com/1032392
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52859}
In the case of an indirect call to an imported function, the target
instance stored in the IFT was actually wrong.
Bug: chromium:834619
Change-Id: Id2ac4158335ecf2b58e1983ce37df852a9ebd1b2
Reviewed-on: https://chromium-review.googlesource.com/1030174
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52831}
WebAssembly.instantiate is polymorphic, it can either take a module
object as parameter, or a buffer source which should be compiled first.
To share code between the two implementations, the module object was
first passed to a promise (i.e. which is the result of compilation).
However, passing the module object to a promise has a side effect if
the module object has a then function. To avoid this side effect I
remove this code sharing and call AsyncInstantiate directly in case
the parameter is a module object.
R=mstarzinger@chromium.org
Bug: chromium:836141
Change-Id: I67b76d0d7761c5aeb2cf1deda45b6842e494eed4
Reviewed-on: https://chromium-review.googlesource.com/1025774
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52755}
On indirect function calls, if the corresponding table entry is empty,
we cannot call {GetCodeFromStartAddress}. In that case, the signature
check will fail anyway, so perform the signature check first, and only
get the code object if the check succeeds.
R=mstarzinger@chromium.org
Bug: chromium:831463
Change-Id: Iead949e4c12502b1a2a3949db2dabab4a184a1e7
Reviewed-on: https://chromium-review.googlesource.com/1005005
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52542}
The bug was fixed in https://crrev.com/c/995796, but this CL adds a
regression test to make sure it stays fixed.
Bug: chromium:827806
Change-Id: I9f4aed364bbd310af4253da457887a8b8015533a
Reviewed-on: https://chromium-review.googlesource.com/993237
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Eric Holk <eholk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52409}
This is a follow-up to https://chromium-review.googlesource.com/981687.
When a wasm function has a large stack frame, the x64 code generator
performs the stack overflow check before constructing the frame. This
requires the use of the `address_of_real_stack_limit` external
reference.
This reference is thread local, so if it is not relocated the stack
overflow check will always fail.
Bug: chromium:808848
Change-Id: I0edf3fe5a006242fc50d0bff44cd9dd0e7d85bd9
Reviewed-on: https://chromium-review.googlesource.com/982906
Commit-Queue: Ben Smith <binji@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52330}
When a wasm function has a large stack frame, the x64 code generator
performs the stack overflow check before constructing the frame. This
requires using the `address_of_real_stack_limit` external reference, as
well as the `ThrowWasmStackOverflow` runtime function.
`ThrowWasmStackOverflow` is called via a generated trampoline, but it is
not a builtin, so the serializer adds it to the `stub_lookup_` map. This
map is encoded by using a monotonically increasing `stub_id` that starts
at 0.
When the function is serialized, a stub is differentiated from a builtin
by which half of the `i32` bits is used, upper or lower. A stub only
uses the lower 16 bits and a builtin only uses the upper 16 bits.
The deserializer checks whether the lower 16 bits are 0; if so, it is
determined to be a builtin. But if the `stub_id` is 0, then it will be
confused with builtin 0 (`RecordWrite`). Calling the builtin instead of
the stub causes a crash.
This CL starts all `stub_id`s at 1, which prevents the builtin/stub
confusion.
There is an additional bug that is not fixed by this CL:
`ThrowWasmStackOverflow` shouldn't be called at all. Currently it is
called because `address_of_real_stack_limit` is a thread-local value
that is not properly relocated.
Bug: chromium:808848
Change-Id: I06b3e650ea58ad717dcc47a3716443e16582e711
Reviewed-on: https://chromium-review.googlesource.com/981687
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Ben Smith <binji@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52252}
When using trap handlers, memory references do not get any checks inserted. This
means there is no check for a null memory as happens when the memory size is
0. Normally this would be correctly caught as an out of bounds access, since the
low memory addresses are not normally mapped. However, if they were mapped for
some reason, we would not catch the out of bounds access.
The fix is to ensure WebAssembly instances always have a guard region even if
the memory is size 0.
This is a rewrite of 5e76ff5a4a
Note that this can lead to a large amount of unnecessary address space usage,
so we share a single reservation for empty array buffers.
Bug: chromium:769637
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ia8e84be6d595e347d3d342959f2c374db1a3f683
Reviewed-on: https://chromium-review.googlesource.com/702657
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Eric Holk <eholk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52163}
On float comparisons, we need a scratch byte register for the setcc
instruction, and if none is available, we spill. But this spilling code
is skipped if one of the operands is NaN. The cache state is updated
however, so following code assumes that the spill happened.
This CL fixes this by spilling before checking for NaN, such that the
spilling code is always executed.
R=titzer@chromium.org
Bug: v8:7582, v8:6600
Change-Id: I768d8de14e494d3ebea181c1f9f3129a4b005396
Reviewed-on: https://chromium-review.googlesource.com/973961
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52162}
See referenced bug: Async compilation can deadlock if a background task
queues the last compilation unit to be finished while the finisher
is already exiting because there was no more work.
This CL fixes this by making the finisher task check for new work after
setting the finisher_is_running_ flag to false.
R=ahaas@chromium.orgCC=kimanh@google.com
Bug: chromium:824681
Change-Id: If1f5700a9fdd5d150b36e37a5d14b692c2b0f3fb
Reviewed-on: https://chromium-review.googlesource.com/975301
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52139}
On ia32, the upper "half stack slot" must be located above the lower
half stack slot (in absolute address), hence the index is
"2 * index - 1" instead of "2 * index + 1". Note that the index
describes the negative offset from the stack pointer.
R=titzer@chromium.org
Bug: v8:7579
Change-Id: If207af405b126ab30043432d7934273e6e2a5330
Reviewed-on: https://chromium-review.googlesource.com/973301
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52116}
There was a bug in spilling i64 constants, in that the half stack slot
*above* should have been filled with the high word instead of the one *below*.
Instead of just fixing this, this CL optimizes spilling x64 constants to the
stack by emitting shorter and faster code, especially if the constant fits in
31 bits (which is the majority of cases).
R=titzer@chromium.org
Bug: v8:7565,v8:6600
Change-Id: Id75ddafe82615930a84333a0c49bd515ccbcc093
Reviewed-on: https://chromium-review.googlesource.com/965062
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51985}
During a C call, a previous value of the stack pointer is stored in a
platform specific callee saved register. Loading the out argument of the
C call might overwrite the value in that register, if the destination
register collides with the platform specific register. Hence, do first
use that register to restore the previous stack pointer, and only then
load the out argument.
Similarly, when pushing arguments to the stack, do first push all
values and then set the platform specific register in order to avoid
overwriting an argument value held in that register.
Drive-by: Fix offset computations for parameters pushed to the stack
for c calls.
R=titzer@chromium.org
Bug: chromium:820802,chromium:820896,chromium:820807,v8:6600
Change-Id: If4567467b7912454f0bd2cad5927233c98894b03
Reviewed-on: https://chromium-review.googlesource.com/959064
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51916}
