Reason for revert:
Failures on waterfall:
e.g. http://build.chromium.org/p/client.v8/builders/V8%20Linux64%20TSAN/builds/11134
Original issue's description:
> [heap] Switch to 500k pages
>
> - Decrease regular heap object size to 300k, keeping the same ration (60%)
> between this limit and page size.
>
> In a follow up, we can now get rid of the new space border page while
> keeping the 1M minimum new space size.
>
> Some results (v8.infinite_scroll; 3 runs):
> - evacuate.avg: +15.3% (1.4->1.2)
> - evacuate.max: +24.4% (2.4->1.8)
>
> BUG=chromium:581412
> LOG=N
> R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
>
> Committed: https://crrev.com/ffe5c670e1559d11e7b252e15fec38765e7dbe4f
> Cr-Commit-Position: refs/heads/master@{#38533}
TBR=hpayer@chromium.org,ulan@chromium.org,yangguo@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:581412
Review-Url: https://codereview.chromium.org/2229403003
Cr-Commit-Position: refs/heads/master@{#38537}
This makes strings_ and values_ in AstValueFactory a linked list through the AstString and AstValue objects. Additionally the CL computes whether strings are convertible to array indexes directly using the AstString's hash + literal bytes just as Name does, rather than indirecting over name if available.
BUG=
Review-Url: https://codereview.chromium.org/2225423002
Cr-Commit-Position: refs/heads/master@{#38536}
LoadInternalField performs some redundant checks in the context of
fast accessors.
This improves the speedup from 20% to 60% for next/previousSibling in
the Dromaeo DOM traversal benchmarks.
BUG=chromium:508898
Review-Url: https://codereview.chromium.org/2186593002
Cr-Commit-Position: refs/heads/master@{#38535}
Now it actually collects free variables of the target function, rather than any reference to a non-stack-allocated variable in any of the inner scopes.
BUG=
Review-Url: https://codereview.chromium.org/2229373002
Cr-Commit-Position: refs/heads/master@{#38534}
- Decrease regular heap object size to 300k, keeping the same ration (60%)
between this limit and page size.
In a follow up, we can now get rid of the new space border page while
keeping the 1M minimum new space size.
Some results (v8.infinite_scroll; 3 runs):
- evacuate.avg: +15.3% (1.4->1.2)
- evacuate.max: +24.4% (2.4->1.8)
BUG=chromium:581412
LOG=N
R=hpayer@chromium.org, ulan@chromium.org, yangguo@chromium.org
Review-Url: https://codereview.chromium.org/2013713003
Cr-Commit-Position: refs/heads/master@{#38533}
This adds an IfSuccess control use in three places:
- in the lowering of isinstance,
- in wasm BuildTrapCode,
- in wasm BuildJavaScriptToNumber,
- WasmGrowMemory is marked as kNoThrow.
BUG=
Review-Url: https://codereview.chromium.org/2226163004
Cr-Commit-Position: refs/heads/master@{#38531}
The scopes of suspended generators can now be accessed through GeneratorMirror
(similar to FrameMirror).
BUG=v8:5235
Review-Url: https://codereview.chromium.org/2228393002
Cr-Commit-Position: refs/heads/master@{#38530}
This makes break point info independent from the code kind being executed.
BUG=v8:5265
Review-Url: https://codereview.chromium.org/2230143002
Cr-Commit-Position: refs/heads/master@{#38528}
Updates a number of AST operations to avoid dereferencing handles
such that they can safely be called off-thread. Also adds a
HandleDereferenceMode argument to some operations where handles are
compared. If handle dereferencing is allowed, the handles are compared
directly, if not then their locations are compared (which relies on the
handles being created in a CanonicalHandleScope).
BUG=v8:5203
TBR=adamk@chromium.org
Review-Url: https://codereview.chromium.org/2223523002
Cr-Commit-Position: refs/heads/master@{#38526}
We now deopt when the lhs of a mod is negative and the rhs is 1 too (previously, we erroneusly returned 0 instead of -0).
BUG=v8:5278
R=bmeurer@chromium.org
Review-Url: https://codereview.chromium.org/2233713002
Cr-Commit-Position: refs/heads/master@{#38525}
Reason for revert:
Found better solution.
Original issue's description:
> [debugger] reapply break points after clearing one shots.
>
> On the debug info, we have a list of existing break points. When we
> step in the debugger, we flood the function with one-shot break points.
> Afterwards, we clear these one-shots by clearing all break locations.
>
> Previously, while clearing break locations, we would skip ones that have
> actual break points. Now we clear all break locations, and then reapply
> break points. This is necessary for the next step, when we encode break
> point info by source position, and not code offset. Encoding by code
> offset would mean that break points are dependent on the code kind we
> use.
>
> R=jgruber@chromium.org
> BUG=v8:5265
>
> Committed: https://crrev.com/808981a22ee19461bbe981ab33b58792d991533e
> Cr-Commit-Position: refs/heads/master@{#38492}
TBR=jgruber@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:5265
Review-Url: https://codereview.chromium.org/2231653003
Cr-Commit-Position: refs/heads/master@{#38521}
Renaming context confuses specialization passes. (And looking through
loop exit renames does not help because we insert Phis for SSA.)
BUG=chromium:635808
Review-Url: https://codereview.chromium.org/2221363002
Cr-Commit-Position: refs/heads/master@{#38517}
We only need these external v8 APIs to expose serialization externally,
so bypassed having an internal layer for now.
BUG=v8:5072
Review-Url: https://codereview.chromium.org/2228073002
Cr-Commit-Position: refs/heads/master@{#38516}
For holey/growing keyed stores, we need to check that there are no
setters in the prototype chain and protect against changes to that
via code dependencies.
R=verwaest@chromium.org
BUG=v8:5275,v8:5276
Review-Url: https://codereview.chromium.org/2231683002
Cr-Commit-Position: refs/heads/master@{#38514}
Implement Bovc and Bnvc instruction macros in macro assembler.
The change 6f920d7d59 revealed an issue with AddBranchOvf for mips R6.
All branching instructions other than BOVC and BNVC are handled by Branch macro, which assures
correct handling of long and short branches. As a consequence, AddBranchOvf for R6 was working correctly
only before trampoline was emitted, but the case of long branch was not handled at all. Implemented
instruction macros enable proper handling of BOVC and BNVC in cases when long branching is needed.
TEST=test/intl/general/case-mapping.js
BUG=
Review-Url: https://codereview.chromium.org/2225323002
Cr-Commit-Position: refs/heads/master@{#38513}
This introduces initial support to handle keyed load access to
String primitives. This is accomplished via the existing operators
StringCharCodeAt and StringFromCharCode, which we already use to
optimize String.prototype.charCodeAt and String.fromCharCode.
R=yangguo@chromium.org
BUG=v8:5267
Review-Url: https://codereview.chromium.org/2232483002
Cr-Commit-Position: refs/heads/master@{#38512}
Rolling v8/build to e1fa1d72e5795ddcaff4ded5b083b1b08e52023a
Rolling v8/third_party/WebKit/Source/platform/inspector_protocol to eaa4c48e31e9f79fb6c4b665dc789dac0f0fb125
Rolling v8/third_party/icu to 53ce631655a61aaaa42b43b4d64abe23e9b8d71f
Rolling v8/tools/mb to d4f0973ff21ab990fed45c3803731c77197d65c8
TBR=machenbach@chromium.org,vogelheim@chromium.org,hablich@chromium.org
Review-Url: https://codereview.chromium.org/2234653002
Cr-Commit-Position: refs/heads/master@{#38511}
Currently we have V8 RuntimeCallStats that is independently from tracing when
running d8 with flag --runtime_call_stats. This patch embeds V8 runtime call
stats into tracing, by having a global table of runtime call counters each
isolate, resetting the table each time we enter a top level trace event, and
dumping the table for each top level trace event. This will make trace file more
compat, as well as enable runtime call stats in tracing system.
This patch adds ~5% overhead to V8 when the category is enabled, we measure the
overhead by running a script when category is enabled.
BUG=v8:5089
Committed: https://crrev.com/d014866173eaa2b548c566217b2c94b1d49385fa
Committed: https://crrev.com/1ca3b73bba4a7253ca8eeef39321d70e7d414331
Committed: https://crrev.com/3f936a5b17754783e92d2146eaf66c88a78ee45b
Review-Url: https://codereview.chromium.org/2187693002
Cr-Original-Original-Original-Commit-Position: refs/heads/master@{#38270}
Cr-Original-Original-Commit-Position: refs/heads/master@{#38314}
Cr-Original-Commit-Position: refs/heads/master@{#38403}
Cr-Commit-Position: refs/heads/master@{#38510}
It can now deal with multiple objects at the same time (but no
aliasing), and it propagates store information upwards across effect
chain splits.
R=jarin
BUG=
Review-Url: https://codereview.chromium.org/2159303002
Cr-Commit-Position: refs/heads/master@{#38509}
This avoids direct scope() accesses and allows us to redirect it over ScopeState.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2226243002
Cr-Commit-Position: refs/heads/master@{#38507}
Block scopes don't need any special treatment here (it's unclear
to me why they ever did). And the has_forced_context_allocation() check
seems, according to our tests, to only have been necessary for proper
handling of 'with' scopes. This patch instead uses the "is_used" bit
to keep track of variables that are accessed from within a with.
R=neis@chromium.org
Review-Url: https://codereview.chromium.org/2220293003
Cr-Commit-Position: refs/heads/master@{#38505}
If a whole scope has forced context allocation, MustAllocateInContext()
already takes care of returning true for all variables, there's no need
to individually call ForceContextAllocation() on them.
R=neis@chromium.org
Review-Url: https://codereview.chromium.org/2222843005
Cr-Commit-Position: refs/heads/master@{#38504}
We assumed we could use a do..while, but the requested amount of slots could've
been 0 and we already started at MIN_CONTEXT_SLOTS. In that case the loop
should've been skipped.
Review-Url: https://codereview.chromium.org/2227723003
Cr-Commit-Position: refs/heads/master@{#38499}
Drive-by-fix: mark isolates as const in stack-allocated scopes
BUG=chromium:630217
Review-Url: https://codereview.chromium.org/2220993003
Cr-Commit-Position: refs/heads/master@{#38496}
On the debug info, we have a list of existing break points. When we
step in the debugger, we flood the function with one-shot break points.
Afterwards, we clear these one-shots by clearing all break locations.
Previously, while clearing break locations, we would skip ones that have
actual break points. Now we clear all break locations, and then reapply
break points. This is necessary for the next step, when we encode break
point info by source position, and not code offset. Encoding by code
offset would mean that break points are dependent on the code kind we
use.
R=jgruber@chromium.org
BUG=v8:5265
Review-Url: https://codereview.chromium.org/2221333002
Cr-Commit-Position: refs/heads/master@{#38492}
When we compile a growing store in TurboFan, we don't pass a (native)
context to the %GrowArrayElements fallback function, as the whole logic
is actually context independent. However, that means that we need to
bailout early in case the object is a prototype, which requires context
dependent checks in the array protector code.
R=cbruni@chromium.org
BUG=chromium:635798
Review-Url: https://codereview.chromium.org/2224253003
Cr-Commit-Position: refs/heads/master@{#38491}
Reason for revert:
crashing on ToT
Original issue's description:
> [heap] Enable black allocation when finalizing incremental marking.
>
> BUG=chromium:630386
>
> Committed: https://crrev.com/bb4e028648a27a6958afc9b5040366b899cda50d
> Cr-Commit-Position: refs/heads/master@{#38487}
TBR=mlippautz@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:630386
Review-Url: https://codereview.chromium.org/2223423003
Cr-Commit-Position: refs/heads/master@{#38490}
