Commit Graph

25749 Commits

Author SHA1 Message Date
hpayer
24622f5f12 Bailout for large object allocations in full code EmitFastOneByteArrayJoin.
BUG=chromium:542823
LOG=n

Review URL: https://codereview.chromium.org/1391373004

Cr-Commit-Position: refs/heads/master@{#31257}
2015-10-14 12:44:45 +00:00
rmcilroy
822b493bee Move AccessorTable from full-codegen.h to ast.h.
It is used by AstGraphBuilder (TF) and BytecodeGenerator (Ignition), so is no
longer a full-codegen datastructure. Removes full-codegen.h dependency from
compiler/ and interpreter/

Review URL: https://codereview.chromium.org/1393393003

Cr-Commit-Position: refs/heads/master@{#31256}
2015-10-14 12:22:22 +00:00
bmeurer
d702f70795 [turbofan] Move native context specialization before inlining.
Perform native context specialization immediately after graph
construction (also after inlinee graph construction). This way
we can do unified inlining before we go to typing and typed
lowering. And we will get better typing due to constants and
(checked) type feedback.

R=mstarzinger@chromium.org
BUG=v8:4470
LOG=n

Review URL: https://codereview.chromium.org/1404123002

Cr-Commit-Position: refs/heads/master@{#31255}
2015-10-14 12:21:03 +00:00
mvstanton
a2a45a1669 Revert of VectorICs: turn on vectors for STORE and KEYED_STORE ics. (patchset #1 id:1 of https://codereview.chromium.org/1396523005/ )
Reason for revert:
Windows crash failure, must be investigated.

Original issue's description:
> VectorICs: turn on vectors for STORE and KEYED_STORE ics.
>
> R=jkummerow@chromium.org
> BUG=
>
> Committed: https://crrev.com/52225f39df578e77b4804506ca4bc15e096f5cab
> Cr-Commit-Position: refs/heads/master@{#31252}

TBR=jkummerow@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/1398123003

Cr-Commit-Position: refs/heads/master@{#31254}
2015-10-14 10:52:32 +00:00
rmcilroy
20fba51253 [Interpreter] Adds Object literal support.
Adds Object literal support to the interpreter. Adds the following bytecodes:
 - ToName
 - CreateObjectLiteral.

BUG=v8:4280
LOG=N

Review URL: https://codereview.chromium.org/1386313005

Cr-Commit-Position: refs/heads/master@{#31253}
2015-10-14 10:10:27 +00:00
mvstanton
52225f39df VectorICs: turn on vectors for STORE and KEYED_STORE ics.
R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1396523005

Cr-Commit-Position: refs/heads/master@{#31252}
2015-10-14 09:27:53 +00:00
mstarzinger
ecc8485da0 [Interpreter] Drop obsolete BuildConstantElements call.
R=rmcilroy@chromium.org

Review URL: https://codereview.chromium.org/1403613003

Cr-Commit-Position: refs/heads/master@{#31251}
2015-10-14 08:59:39 +00:00
machenbach
749b419a82 [swarming] Isolate test262.
BUG=chromium:535160
LOG=n

Review URL: https://codereview.chromium.org/1402073003

Cr-Commit-Position: refs/heads/master@{#31250}
2015-10-14 08:44:27 +00:00
mstarzinger
8ff6a0c005 [turbofan] First rudimentary inlining heuristic.
This is a first prototype for a rudimentary inlining heuristic allowing
enabling of general inlining based existing budget flags. Also note that
this approach does not yet work for multi-level inlining, for now the
list of candidates is processed exactly once.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1406543002

Cr-Commit-Position: refs/heads/master@{#31249}
2015-10-14 08:37:19 +00:00
machenbach
6628b77d57 [test] Mark intl test as failing on linux.
BUG=v8:2899
LOG=n
TBR=jochen@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1396293006

Cr-Commit-Position: refs/heads/master@{#31248}
2015-10-14 07:09:36 +00:00
bmeurer
1262a41c1f [turbofan] Disable native context specialization for the asm.js pipeline.
R=jarin@chromium.org
BUG=v8:4470
LOG=n

Review URL: https://codereview.chromium.org/1401923004

Cr-Commit-Position: refs/heads/master@{#31247}
2015-10-14 05:51:24 +00:00
dpranke
dd8206c81a Use //build_overrides instead of //build/modules for GN builds.
R=machenbach@chromium.org
BUG=chromium:541791
LOG=N

Review URL: https://codereview.chromium.org/1396743005

Cr-Commit-Position: refs/heads/master@{#31246}
2015-10-13 23:56:25 +00:00
akos.palfi
d3f97a972f MIPS: Fix movz_movn test failure with Clang.
Correctly initialize the high FP registers.

TEST=test-assembler-mips/movz_movn,test-assembler-mips64/movz_movn
BUG=

Review URL: https://codereview.chromium.org/1399413003

Cr-Commit-Position: refs/heads/master@{#31245}
2015-10-13 22:16:16 +00:00
hpayer
c1a81536ed Do not allow large object allocation from optimized code.
BUG=

Review URL: https://codereview.chromium.org/1406593002

Cr-Commit-Position: refs/heads/master@{#31244}
2015-10-13 19:20:19 +00:00
Djordje.Pesic
d6741e80ea MIPS: Fix mina in mips32 and mips64 simulator
Fix mina instruction in mips32 and mips64 simulator according to IEEE 754-2008 standard

BUG=

Review URL: https://codereview.chromium.org/1402923002

Cr-Commit-Position: refs/heads/master@{#31243}
2015-10-13 15:47:07 +00:00
mvstanton
f2f46aff8b VectorICs: use a vector slot to aid in array literal processing.
The lack of a vector slot for the keyed store operation in filling in
non-constant array literal properties led to undesirable contortions in
compilers downwind of full-codegen. The use of a single slot to initialize all
the array elements is sufficient.

BUG=

Review URL: https://codereview.chromium.org/1405503002

Cr-Commit-Position: refs/heads/master@{#31242}
2015-10-13 14:48:49 +00:00
jkummerow
c706c0fa19 Add C++ implementation of Object.defineProperties
Not used yet, so this CL shouldn't change behavior.

Review URL: https://codereview.chromium.org/1368753003

Cr-Commit-Position: refs/heads/master@{#31241}
2015-10-13 14:38:35 +00:00
rmcilroy
6a10a9af3b [Interpreter] Add array literal support.
Adds array literal support to the interpreter. Currently constructed
array elements don't have type feedback slots, so also adds support for
generic keyed store operations.

Adds the following bytecodes:
 - CreateArrayLiteral
 - KeyedStoreICGeneric

BUG=v8:4280
LOG=N

Review URL: https://codereview.chromium.org/1400753003

Cr-Commit-Position: refs/heads/master@{#31240}
2015-10-13 14:00:57 +00:00
hpayer
1f98ef6ec8 Revert of [heap] Reland decrease large object limit for regular heap objects. (patchset #3 id:40001 of https://codereview.chromium.org/1401173003/ )
Reason for revert:
Still one failing test on Windows.

Original issue's description:
> [heap] Reland decrease large object limit for regular heap objects.
>
> BUG=
>
> Committed: https://crrev.com/9b91bf3a630b5382080ba1d7f9f3e88de3de8a95
> Cr-Commit-Position: refs/heads/master@{#31228}

TBR=mlippautz@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/1395653003

Cr-Commit-Position: refs/heads/master@{#31239}
2015-10-13 13:47:23 +00:00
rmcilroy
c0185b7d98 [Interpreter] Add support for new local function context creation.
Adds support for creation of new local function contexts (or script context for
top-level code). As part of this, also adds support for context push/pop
operations using a ContextScope object in BytecodeGenerator. Adds the following
bytecodes:
 - PushContext
 - PopContext

Support for inner contexts and loading from / storing to context allocated
variables will come in a future CL.

BUG=v8:4280
LOG=N

Review URL: https://codereview.chromium.org/1379793004

Cr-Commit-Position: refs/heads/master@{#31238}
2015-10-13 13:09:56 +00:00
ofrobots
010897c16a Reland improve perf_basic_prof filename reporting
Using perf-basic-prof in the test-case was problematic on windows. Use
CodeEventLogger directly.

Previous issue: https://codereview.chromium.org/1396843004/

R=jkummerow@chromium.org,yangguo@chromium.org,yurys@chromium.org
BUG=chromium:539892
LOG=N

Committed: https://crrev.com/701ba0b255f9c34f4b8c43584ef1e35040474e7d
Cr-Commit-Position: refs/heads/master@{#31197}

patch from issue 1396843004 at patchset 60001 (http://crrev.com/1396843004#ps60001)

Review URL: https://codereview.chromium.org/1403763002

Cr-Commit-Position: refs/heads/master@{#31237}
2015-10-13 12:45:31 +00:00
yangguo
4f7caf2f57 Debugger: fix stepping when break points are deactivated.
The flag for deactivating break points also affects stepping, since both
are implemented via debug break slots. Fixing this by introducing a new
flag solely responsible for deactivating actual break points.

R=mvstanton@chromium.org
BUG=chromium:119800
LOG=N

Review URL: https://codereview.chromium.org/1402913002

Cr-Commit-Position: refs/heads/master@{#31236}
2015-10-13 12:44:28 +00:00
mlippautz
ec1046f9f8 Revert of [heap] Divide available memory upon compaction tasks (patchset #9 id:340001 of https://codereview.chromium.org/1382003002/ )
Reason for revert:
Failing tests: https://chromegw.corp.google.com/i/client.v8/builders/V8%20Linux%20-%20arm64%20-%20sim%20-%20nosnap%20-%20debug%20-%202/builds/3804/steps/Check/logs/DontLeakGlobalObjects

Original issue's description:
> [heap] Divide available memory upon compaction tasks
>
> - Fairly (round-robin) divide available memory upon compaction tasks.
> - Ensure an upper limit (of memory) since dividing is O(n) for n free-space
>   nodes.
> - Refill from free lists managed by sweeper once a compaction space becomes
>   empty.
>
> Assumption for dividing memory: Memory in the free lists is sparse upon starting
> compaction (which means that only few nodes are available), except for memory
> reducer GCs, which happen in idle time though (so it's less of a problem).
>
> BUG=chromium:524425
> LOG=N
>
> Committed: https://crrev.com/30236c052ba9266fc55412a8fd63b17f683ff40b
> Cr-Commit-Position: refs/heads/master@{#31234}

TBR=ulan@chromium.org,hpayer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:524425

Review URL: https://codereview.chromium.org/1406533002

Cr-Commit-Position: refs/heads/master@{#31235}
2015-10-13 12:00:11 +00:00
mlippautz
30236c052b [heap] Divide available memory upon compaction tasks
- Fairly (round-robin) divide available memory upon compaction tasks.
- Ensure an upper limit (of memory) since dividing is O(n) for n free-space
  nodes.
- Refill from free lists managed by sweeper once a compaction space becomes
  empty.

Assumption for dividing memory: Memory in the free lists is sparse upon starting
compaction (which means that only few nodes are available), except for memory
reducer GCs, which happen in idle time though (so it's less of a problem).

BUG=chromium:524425
LOG=N

Review URL: https://codereview.chromium.org/1382003002

Cr-Commit-Position: refs/heads/master@{#31234}
2015-10-13 11:05:18 +00:00
yangguo
82b30828bd Check for validity when accessing call site objects in runtime.
R=jkummerow@chromium.org
BUG=chromium:528379
LOG=N

Review URL: https://codereview.chromium.org/1404613002

Cr-Commit-Position: refs/heads/master@{#31233}
2015-10-13 10:53:22 +00:00
mstarzinger
2633401137 Workaround for borked sem_init library function.
This adds a workaround that zeroes out semaphores before they are
initialized. Some versions of sem_init (e.g. GLIBC_2.0) fail to fully
zero out the semaphore, leading to {errno == ENOSYS} with subsequent
sem_timedwait calls.

R=machenbach@chromium.org
BUG=chromium:536813
LOG=n

Review URL: https://codereview.chromium.org/1407463002

Cr-Commit-Position: refs/heads/master@{#31232}
2015-10-13 09:53:10 +00:00
rmcilroy
66e5937337 [Interpreter] Add function literal support.
Adds function literal support and add support for OTHER_CALLS which can be
made when calling a function literal.

Adds the CreateClosure bytecode.

BUG=v8:4280
LOG=N

Review URL: https://codereview.chromium.org/1396693003

Cr-Commit-Position: refs/heads/master@{#31231}
2015-10-13 09:40:01 +00:00
mstarzinger
e5320d8a59 Move builtin JavaScript sources into own directory.
This moves JavaScript source files that are bundled with V8 into a
separate directory. The goal is to improve code readability and also
being able to formalize ideal reviewers by subsequently adding the
OWNERS file. These files almost exclusively contain implementations
of methods fully specified by ES6.

Note that files in the "debug" directory as well as the "d8.js" file
aren't affected by this change.

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/1398733002

Cr-Commit-Position: refs/heads/master@{#31230}
2015-10-13 09:38:26 +00:00
yangguo
1a94bc20a1 Fix Error object value lookups.
Looking up 'name' and 'message' properties at the same time and loading
the properties later can cause assertion failure if one of the properties
is an accessor and calling it changes the holder map. That may invalidate
the other lookup.

R=jkummerow@chromium.org
BUG=chromium:542101
LOG=N

Review URL: https://codereview.chromium.org/1403923002

Cr-Commit-Position: refs/heads/master@{#31229}
2015-10-13 09:26:47 +00:00
hpayer
9b91bf3a63 [heap] Reland decrease large object limit for regular heap objects.
BUG=

Review URL: https://codereview.chromium.org/1401173003

Cr-Commit-Position: refs/heads/master@{#31228}
2015-10-13 08:43:21 +00:00
jacob.bramley
1d9d7821ff [arm64] Fix a string test in StringCompareStub.
This fixes a regression in Octane's pdf.js after r30818. The bug itself
has been present for a long time, but StringCompareStub wasn't actually
used until r30818 so it went unnoticed.

Review URL: https://codereview.chromium.org/1399983002

Cr-Commit-Position: refs/heads/master@{#31227}
2015-10-13 08:40:42 +00:00
yangguo
8e4ceb613d Debugger: for frame details, use callee context to wrap receiver.
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1406493002

Cr-Commit-Position: refs/heads/master@{#31226}
2015-10-13 07:58:15 +00:00
mstarzinger
b79f397f41 [turbofan] Base JSCreateClosure comparison on location.
This changes hashing and comparison functions for JSCreateClosure
operators to be based in the handle location instead of the referenced
object identity. This is in sync with all other JS operators.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1388233007

Cr-Commit-Position: refs/heads/master@{#31225}
2015-10-13 07:43:43 +00:00
mtrofin
efdcd20267 [turbofan] Splinter into one range.
Before this CL, we created one live range per successive set of
deferred blocks. For scenarios with many such blocks, this creates
an upfront pressure for the register allocator to deal with many ranges.
Linear sorts ranges, which is a super-linear operation.

The change places all deferred intervals into one range, meaning that,
at most, there will be twice as many live ranges as the original set. In
pathological cases (benchmarks/Compile/slow_nbody1.js), this change
halves the compilation time. We see some improvements elsewhere,
notably SQLite at ~4-5%.

We may be able to avoid the subsequent merge. Its cost is the
additional ranges it may need to create. The sole reason for the merge
phase is to provide an unchanged view of the world to the subsequent
phases. With the at-most-one splinter model, we may be able to teach
the other phases about splintering - should we find perf hindrances
due to merging.

Review URL: https://codereview.chromium.org/1391023007

Cr-Commit-Position: refs/heads/master@{#31224}
2015-10-13 03:58:19 +00:00
v8-autoroll
6327ed0c92 Update V8 DEPS.
Rolling v8/tools/clang to a56c9063d582660d29b32f967ce20875c589fc3e

TBR=machenbach@chromium.org,vogelheim@chromium.org,hablich@chromium.org

Review URL: https://codereview.chromium.org/1407453002

Cr-Commit-Position: refs/heads/master@{#31223}
2015-10-13 03:23:09 +00:00
dusan.m.milosavljevic
0c5c479574 MIPS: [turbofan] Improve test and equality compare with zero and constants.
TEST=
BUG=

Review URL: https://codereview.chromium.org/1400833002

Cr-Commit-Position: refs/heads/master@{#31222}
2015-10-13 00:15:48 +00:00
mbrandy
32f13dfbec PPC: Refine "Test for var declarations in eval which conflict with let".
R=joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com, dstence@us.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1390953004

Cr-Commit-Position: refs/heads/master@{#31221}
2015-10-12 19:09:35 +00:00
mtrofin
cbbfc0ff87 [turbofan] More efficient splintering.
For live ranges with many use positions, such as those encountered in
some unity asm.js code, this change significantly reduces compile time
(e.g. benchmarks/Compile/slow_nbody1.js: from ~6s to 2s). The
improvement is solely due to regressions (fixed by this CL) due to
splintering.

This CL does not fully address compile time problems for large
functions in Turbofan, but constitutes a step in the right direction.

Review URL: https://codereview.chromium.org/1386253004

Cr-Commit-Position: refs/heads/master@{#31220}
2015-10-12 18:34:40 +00:00
hpayer
9790e494a4 Revert of [heap] Decrease large object limit for regular heap objects. (patchset #3 id:40001 of https://codereview.chromium.org/1392823003/ )
Reason for revert:
broken tests

Original issue's description:
> [heap] Decrease large object limit for regular heap objects.
>
> Committed: https://crrev.com/e887d42342b3449f64fb6078f84f6c55bae3978d
> Cr-Commit-Position: refs/heads/master@{#31217}

TBR=mlippautz@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1402823002

Cr-Commit-Position: refs/heads/master@{#31219}
2015-10-12 17:11:24 +00:00
adamk
18534dffc9 Don't throw on assignment to function name binding in harmony sloppy mode
BUG=v8:4482
LOG=n

Review URL: https://codereview.chromium.org/1397513004

Cr-Commit-Position: refs/heads/master@{#31218}
2015-10-12 16:55:35 +00:00
hpayer
e887d42342 [heap] Decrease large object limit for regular heap objects.
Review URL: https://codereview.chromium.org/1392823003

Cr-Commit-Position: refs/heads/master@{#31217}
2015-10-12 16:39:33 +00:00
mvstanton
76bee80672 Remove FLAG_cache_optimized_code because we always have it on.
The data structure that holds the optimized code is becoming essential for
additional functionality, too.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1399033002

Cr-Commit-Position: refs/heads/master@{#31216}
2015-10-12 16:10:59 +00:00
mstarzinger
c84c96caa2 Remove cycle breaking include from mksnapshot.cc file.
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/1399623004

Cr-Commit-Position: refs/heads/master@{#31215}
2015-10-12 15:21:26 +00:00
mstarzinger
e494bcca7e Reland "[heap] No leakage of mark-compact.h outside of heap."
The revert of the original change in question did not recover
regressions, which allows us to reland this cleanup after all.

This reverts commit 871529b4d8.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/1393773004

Cr-Commit-Position: refs/heads/master@{#31214}
2015-10-12 15:20:31 +00:00
mbrandy
f7a35b9a6e PPC: Vector ICs: Get rid of stack arguments on ia32 transitioning stores.
Port 2d4aeaad2f

Original commit message:
    The stack manipulation was expensive. Two virtual registers are better.

R=mvstanton@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com, dstence@us.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1390343004

Cr-Commit-Position: refs/heads/master@{#31213}
2015-10-12 15:18:32 +00:00
machenbach
fe6ff6523e Revert of Don't compile functions in a context the caller doesn't have access to (patchset #1 id:1 of https://codereview.chromium.org/1393713006/ )
Reason for revert:
[Sheriff] Breaks layout tests. Please add needsmanualrebaseline upstream first if intended. E.g.:
http://build.chromium.org/p/client.v8.fyi/builders/V8-Blink%20Linux%2064/builds/2263

(one of them is a timeout that only happens with this commit)

Original issue's description:
> Don't compile functions in a context the caller doesn't have access to
>
> Instead, just return undefined.
>
> A side effect of this is that it's no longer possible to compile
> functions in a detached context.
>
> Based on https://codereview.chromium.org/294073002 but taking access
> check callbacks into account
>
> BUG=chromium:541703
> R=verwaest@chromium.org
> LOG=y
>
> Committed: https://crrev.com/9a5e2f512c4aa90563eb575605c2a8c2a92ac9f4
> Cr-Commit-Position: refs/heads/master@{#31208}

TBR=verwaest@chromium.org,jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:541703

Review URL: https://codereview.chromium.org/1395583004

Cr-Commit-Position: refs/heads/master@{#31212}
2015-10-12 14:36:00 +00:00
littledan
d515e5138d Test for var declarations in eval which conflict with let
Previously, name conflicts between var and let declarations were only
made into exceptions if they were visible at parse-time. This patch adds
runtime checks so that sloppy-mode direct eval can't introduce conflicting
var declarations. The change is implemented by traversing the scope chain
when a direct eval introduces a var declaration to look for conflicting
let declarations, up to the function boundary.

BUG=v8:4454
R=adamk
LOG=Y

Review URL: https://codereview.chromium.org/1382513003

Cr-Commit-Position: refs/heads/master@{#31211}
2015-10-12 14:31:01 +00:00
mythria
9feb530594 Adds support for following operators
-Bitwise Or
  -Bitwise Xor
  -Bitwise And

Adds the above bytecodes, support to BytecodeGenerator and BytecodeArrayBuilder to enable it's use, it's implementation and tests.

BUG=v8:4280
LOG=N

Review URL: https://codereview.chromium.org/1386133002

Cr-Commit-Position: refs/heads/master@{#31210}
2015-10-12 13:36:07 +00:00
neis
f113f3c384 [es6] Stage --harmony-completion.
R=rossberg
BUG=

Review URL: https://codereview.chromium.org/1403683003

Cr-Commit-Position: refs/heads/master@{#31209}
2015-10-12 12:36:47 +00:00
jochen
9a5e2f512c Don't compile functions in a context the caller doesn't have access to
Instead, just return undefined.

A side effect of this is that it's no longer possible to compile
functions in a detached context.

Based on https://codereview.chromium.org/294073002 but taking access
check callbacks into account

BUG=chromium:541703
R=verwaest@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/1393713006

Cr-Commit-Position: refs/heads/master@{#31208}
2015-10-12 12:23:18 +00:00