This is just a very minor cleanup avoiding unnecessary slow stub usage.
BUG=v8:5561
Change-Id: Ib145606520b5b7019f1e55d3aec7caa956c7e090
Reviewed-on: https://chromium-review.googlesource.com/457337
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43989}
On S390, GCC seems to treat size_t as 'long_unsigned_int'. Which conflicts with the %d format specifier since that expects an 'int'. So changing the macro to PRIuS. This seems to fix the issue while retaining the expected behaviour for all other systems.
R=mlippautz@chromium.org, hpayer@chromium.org
BUG=none
Review-Url: https://codereview.chromium.org/2759963003
Cr-Commit-Position: refs/heads/master@{#43988}
Rebase of https://chromium-review.googlesource.com/446336 was broken.
Didn't fail any tests because the feature is behind a flag.
BUG=
Change-Id: I6009bfac27aa896e20446f002f5db707aab7e34c
Reviewed-on: https://chromium-review.googlesource.com/457040
Commit-Queue: Marja Hölttä <marja@chromium.org>
Reviewed-by: Daniel Vogelheim <vogelheim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43984}
Inspired by bsuite/compare-baseline.py and tools/perf-to-html.py. The main difference
from the latter is that the new script does not require json files preprocessing and it
is able to compare multiple json files at once (just like bsuite/compare-baseline.py).
This CL also makes tools/run_perf.py relative path friendly.
BUG=
Change-Id: I7be796fa920f8360703806f32767b992eec0f9c2
Reviewed-on: https://chromium-review.googlesource.com/456286
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43982}
Reason for revert:
Breaks layout tests:
https://build.chromium.org/p/client.v8.fyi/builders/V8-Blink%20Linux%2064/builds/14312
See https://github.com/v8/v8/wiki/Blink-layout-tests
Original issue's description:
> [wasm] Transferrable modules
>
> We want to restrict structured cloning in Chrome to:
> - postMessage senders and receivers that are co-located
> in the same process
> - indexedDB (just https).
>
> For context, on the Chrome side, we will achieve the postMessage part
> by using a mechanism similar to transferrables: the
> SerializedScriptValue will have a list of wasm modules, separate from
> the serialized data stream; and this list won't be copied cross
> process boundaries. The IDB part is achieved by explicitly opting in
> reading/writing to the serialization stream. To block attack vectors
> in IPC cases, the default for deserialization will be to expect data
> in the wasm transfers list.
>
> This change is the V8 side necessary to enabling this design. We
> introduce TransferrableModule, an opaque datatype exposed to the
> embedder. Internally, TransferrableModules are just serialized data,
> because we don't have a better mechanism, at the moment, for
> de-contextualizing/re-contextualizing wasm modules (wrt Isolate and
> Context).
>
> The chrome defaults will be implemented in the
> serialization/deserialization delegates on that side. For the v8 side
> of things, in the absence of a serialization delegate, the V8
> serializer will write to serialization stream. In the absence of a
> deserialization delegate, the deserializer won't work. This asymmetry
> is intentional - it communicates to the embedder the need to make a
> policy decision, otherwise wasm serialization/deserialization won't
> work "out of the box".
>
> BUG=v8:6079
>
> Review-Url: https://codereview.chromium.org/2748473004
> Cr-Commit-Position: refs/heads/master@{#43955}
> Committed: 99743ad460TBR=jbroman@chromium.org,bradnelson@chromium.org,mtrofin@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:6079
Review-Url: https://codereview.chromium.org/2762163002
Cr-Commit-Position: refs/heads/master@{#43981}
This frees up the InterpreterAssembler for no longer being linked
into the main library.
BUG=v8:6055
Review-Url: https://codereview.chromium.org/2759093004
Cr-Commit-Position: refs/heads/master@{#43979}
For MIPS32, instructions ldc1 and sdc1 are moved into macro-assembler
and renamed as Ldc1 and Sdc1. The reason for placing them into
macro-assembler is that they emmit two or three instructions.
TEST=test/cctest/test-assembler-mips,
test/cctest/test-code-stubs-mips,
test/cctest/test-macro-assembler-mips
BUG=
Review-Url: https://codereview.chromium.org/2751973002
Cr-Commit-Position: refs/heads/master@{#43977}
This CL makes the interpreter reentrant by allowing different
activations to be live at the same time. The wasm interpreter keeps a
list of activations and stores the stack height at the start of each
activation. This information is used to unwind just one activation, or
show the right portion of the interpreter stack for each interpreter
entry frame.
The WasmDebugInfo object stores a mapping from frame pointer (of the
interpreter entry) to the activation id in order to identify the
activation based on the physical interpreter entry frame.
R=titzer@chromium.org, ahaas@chromium.org
BUG=v8:5822
Change-Id: Ibbf93f077f907213173a92e0a2f7f3556515e8eb
Reviewed-on: https://chromium-review.googlesource.com/453958
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43976}
In the tick processor, in cases where there are a lot of ticks (e.g.
long running programs), JSON.stringify could throw a range exception
because the created string is too large.
Instead of creating the entire JSON string in memory, we now write the
top-level parts of the JSON manually, writing out the ticks individually
instead of all together.
Review-Url: https://codereview.chromium.org/2754683002
Cr-Commit-Position: refs/heads/master@{#43973}
The called runtime function never returns, thus we don't need to emit
return code afterwards.
R=ahaas@chromium.org
Change-Id: I4adb5492b1d5bcb8f644f9544137e07196ac61e4
Reviewed-on: https://chromium-review.googlesource.com/456507
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43972}
For CheckBounds(index,length) we know that the length must be in
Unsigned31 range. Thus there's no observable difference for index
values in the range [-2^31,-1] and the range [2^31,2^32-1], both
are considered out-of-bounds; also it's safe to truncate -0 to 0
wrt. CheckBounds. Thus we can safely pass Word32 truncation if the
index is in Integral32 \/ MinusZero. Usually this generates the same
code, but some index computations can benefit from the Word32 truncation
and avoid going to double because the result would be outside the valid
Signed32 or Unsigned32 ranges.
R=jarin@chromium.org
BUG=v8:5267
Review-Url: https://codereview.chromium.org/2760213003
Cr-Commit-Position: refs/heads/master@{#43969}
When returning from the runtime function, move jssp back to csp. The
csp might have been changed by the runtime function, but jssp should
have been restored to its original value.
R=ahaas@chromium.org
BUG=v8:5822
Change-Id: I300263a586ca546a4d7f925730f1f38b680379ca
Reviewed-on: https://chromium-review.googlesource.com/457372
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43967}
We used to clear invocation counts when enabling precise coverage.
This is not necessary, and we could continue to use the existing
invocation counts on the heap. The old behavior can be achieved
by explicitly resetting the counts by polling coverage data.
R=jgruber@chromium.org,caseq@chromium.org
BUG=v8:5808
Review-Url: https://codereview.chromium.org/2768453002
Cr-Commit-Position: refs/heads/master@{#43964}
The flag is already on by default
R=clemensh@chromium.org
Change-Id: Ie4ede8191336a102cab9d7f972a3d10a15d1a54d
Reviewed-on: https://chromium-review.googlesource.com/456287
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43963}
Also move phi NodeVector in TryCloneBranch to temporary zone.
BUG=chromium:700364
Change-Id: Id19d51dae63ed5a6f5dccbba77a19b3663fd325e
Reviewed-on: https://chromium-review.googlesource.com/456285
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43962}
Improve visibility within the distributed wasm team.
Created wasm watchlist, and added wasm-team - the union of
MUC and MTV wasm teams.
BUG=
Review-Url: https://codereview.chromium.org/2759053002
Cr-Commit-Position: refs/heads/master@{#43959}
When we hit a call to String.prototype.concat builtin, where we can
infer that the receiver is a String and there's exactly one parameter,
which is of type PlainPrimitive, then we can reduce that to a call to
the StringAddStub instead, optionally converting the non-String - but
PlainPrimitive - parameter to a String.
BUG=v8:5267
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2758383002
Cr-Commit-Position: refs/heads/master@{#43957}
Collect type information of return values.
Use *one* feedback slot per function for all its return
statements. For assignments, we currently use
several slots per function, because not all
assignments refer to the same variable.
Instead of the variable names, pass the
source location and print the function name.
Add an integration test for --type-profile that checks
for crashes.
Remove type feedback for assignments for now as it convolutes the output.
************ Function with 2 return statements ********
function testFunction(param, flag) {
// We want to test 2 different return positions in one function.
if (flag) {
var first_var = param;
return first_var;
}
var second_var = param;
return second_var;
}
testFunction({});
testFunction(123, true);
testFunction('hello');
testFunction(undefined);
*******************************************************
************* Sample Output ***************************
Function: testFunction
424: Object
374: number
424: string
424: undefined
*******************************************************
Missing work:
* Handle fall-off returns
* Collect types for parameters
* Remove duplicates from the list of collected types and use a common base class.
BUG=v8:5935
Review-Url: https://codereview.chromium.org/2755973002
Cr-Commit-Position: refs/heads/master@{#43956}
We want to restrict structured cloning in Chrome to:
- postMessage senders and receivers that are co-located
in the same process
- indexedDB (just https).
For context, on the Chrome side, we will achieve the postMessage part
by using a mechanism similar to transferrables: the
SerializedScriptValue will have a list of wasm modules, separate from
the serialized data stream; and this list won't be copied cross
process boundaries. The IDB part is achieved by explicitly opting in
reading/writing to the serialization stream. To block attack vectors
in IPC cases, the default for deserialization will be to expect data
in the wasm transfers list.
This change is the V8 side necessary to enabling this design. We
introduce TransferrableModule, an opaque datatype exposed to the
embedder. Internally, TransferrableModules are just serialized data,
because we don't have a better mechanism, at the moment, for
de-contextualizing/re-contextualizing wasm modules (wrt Isolate and
Context).
The chrome defaults will be implemented in the
serialization/deserialization delegates on that side. For the v8 side
of things, in the absence of a serialization delegate, the V8
serializer will write to serialization stream. In the absence of a
deserialization delegate, the deserializer won't work. This asymmetry
is intentional - it communicates to the embedder the need to make a
policy decision, otherwise wasm serialization/deserialization won't
work "out of the box".
BUG=v8:6079
Review-Url: https://codereview.chromium.org/2748473004
Cr-Commit-Position: refs/heads/master@{#43955}
The validate-asm flag now implies lazy compilation.
R=titzer@chromium.org, ahaas@chromium.org
BUG=v8:5991
Change-Id: I00fb5ddbe13440941a3fafd9175cc9a5d182e15a
Reviewed-on: https://chromium-review.googlesource.com/451318
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43952}
This reverts commit ea4346be7f.
Reason for revert: This causes a crash in a Node.js test (parallel/test-repl-options). I can reproduce the crash locally. At a first glance, it doesn't look like we're doing anything terribly wrong in Node that explains the crash. I'll have a closer lock tomorrow. Feel free to reland if needed, we can always deactivate the test.
Original change's description:
> [ic] Migrate StoreGlobal to data handler
>
> BUG=v8:5561
>
> Change-Id: If4c679c97af199ce1c90d055627186123bc88574
> Reviewed-on: https://chromium-review.googlesource.com/456698
> Reviewed-by: Igor Sheludko <ishell@chromium.org>
> Commit-Queue: Toon Verwaest <verwaest@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#43944}
TBR=ishell@chromium.org,verwaest@chromium.org,v8-reviews@googlegroups.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:5561
Change-Id: I790ff9ab45016749fe2f3982045f497a995e282e
Reviewed-on: https://chromium-review.googlesource.com/456505
Reviewed-by: Franziska Hinkelmann <franzih@chromium.org>
Commit-Queue: Franziska Hinkelmann <franzih@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43951}
as InterpreterGenerator. This is in preparation for no longer including
the bytecode handler generation code in the main library.
BUG=v8:6055
Review-Url: https://codereview.chromium.org/2765433003
Cr-Commit-Position: refs/heads/master@{#43949}
When instantiating the wasm interpreter, pass the start address of the
global variables. This was nullptr before, leading to a crash if
debugging a program which accesses globals.
With test.
R=ahaas@chromium.org, titzer@chromium.org
BUG=v8:5822
Change-Id: I5f419790042ef9a00787df093a07e5e5835d55bd
Reviewed-on: https://chromium-review.googlesource.com/456219
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43948}
Before, we were redirecting each function to the interpreter by iterating all
code and patching all call sites using this one function. The runtime was
hence quadratic if all functions were redirected to the interpreter as
done by the --wasm-interpret-all flag.
This CL fixes this to only iterate the code once and redirecting an
arbitrary number of function.
R=ahaas@chromium.org, titzer@chromium.org
BUG=v8:5822
Change-Id: Ia4f2e94a2468f9bef3035b599e1f8a18acf309da
Reviewed-on: https://chromium-review.googlesource.com/455785
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43946}
Defers a label in the LoadIC fast-path which was doing a call that was
forcing the entire LoadIC fast-path to have a frame. The label was
introduced in https://chromium-review.googlesource.com/c/455858/.
Change-Id: Icc8f7243c133cfa0ad60ede0d0f5651b639634e9
Reviewed-on: https://chromium-review.googlesource.com/456504
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43945}
This fixes a bug where an exported function is being specialized, but
the callsite inside the JS_TO_WASM function was patched to call an
interpreter entry instead. We would not identify the call site as the
one to be patched during specialization, and would thus fail a DCHECK.
R=ahaas@chromium.org
BUG=v8:5822, chromium:702839
Change-Id: I148d98333051c399a4cb11bd9620b396f4eb261d
Reviewed-on: https://chromium-review.googlesource.com/456282
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43942}
Currently the incremental marking visitor treats elements of normalized
map caches weakly by coloring the caches grey without pusing to marking
deque.
The mark-compact prologue then clears all normalized map caches.
We can achieve similar effect by just clearing the caches in the marking
visitor.
BUG=chromium:694255
Review-Url: https://codereview.chromium.org/2745183002
Cr-Commit-Position: refs/heads/master@{#43941}
Makes disassembly jump target printing look more like the output of
objdump, for compatibility with perf's jump arrows. This includes
swapping the order of address and offset, and making the offset and line
numbers hex.
As a drive-by, print comment lines in objdump-v8 so that they can be
shown/hidden as "source" lines by perf.
Review-Url: https://codereview.chromium.org/2757263002
Cr-Commit-Position: refs/heads/master@{#43940}
