Commit Graph

28 Commits

Author SHA1 Message Date
titzer
41795b8ae2 [turbofan] Add bounds check to Node::InputAt(index) and fix tests that go out of bounds.
BUG=

Review URL: https://codereview.chromium.org/1149563004

Cr-Commit-Position: refs/heads/master@{#28540}
2015-05-21 10:59:40 +00:00
Benedikt Meurer
2d827809e0 [turbofan] Add language mode to JSCallFunction operator.
Also do some drive-by-cleanup to the unittests.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1107883002

Cr-Commit-Position: refs/heads/master@{#28066}
2015-04-27 10:44:30 +00:00
bmeurer
ff89876bb9 [turbofan] Fix lowering of Math.max for integral inputs.
R=jarin@chromium.org
BUG=chromium:468162
LOG=y

Review URL: https://codereview.chromium.org/1027753002

Cr-Commit-Position: refs/heads/master@{#27341}
2015-03-20 12:05:19 +00:00
bmeurer
83f157bc18 [turbofan] Use builtin inlining mechanism for Math.abs and Math.sqrt.
Reduces the amount of custom support code for Math functions in TurboFan
and allows for more general inlining (i.e. independent of parameter
types).

BUG=v8:3952
LOG=n
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/1004083002

Cr-Commit-Position: refs/heads/master@{#27172}
2015-03-13 07:06:15 +00:00
bmeurer
1982186b6f [turbofan] Use builtin inlining mechanism for Math.floor.
BUG=v8:3952
LOG=n
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/997513002

Cr-Commit-Position: refs/heads/master@{#27098}
2015-03-10 11:59:31 +00:00
bmeurer
022ea7e057 [turbofan] Unify Math.floor / Math.ceil optimization.
Provide an intrinsic %MathFloor / %_MathFloor that is used to optimize
both Math.ceil and Math.floor, and use the JS inlining mechanism to
inline Math.ceil into TurboFan code. Although we need to touch code
outside of TurboFan to make this work, this does not affect the way we
handle Math.ceil and/or Math.floor in CrankShaft, because for CrankShaft
the old-style builtin function id based inlining still kicks in first.

Once this solution is stabilized, we can use it for Math.floor as well.
And once that is settled, we can establish it as the unified way to
inline builtins, and get rid of the specialized builtin function id
based inlining at some point.

Note that "builtin" applies to basically every piece of internal
JavaScript/intrinsics based code, so this also applies to the yet to be
defined JavaScript based code stubs and handlers.

BUG=v8:3953
LOG=n
R=yangguo@chromium.org,svenpanne@chromium.org

Review URL: https://codereview.chromium.org/990963003

Cr-Commit-Position: refs/heads/master@{#27086}
2015-03-10 08:42:53 +00:00
bmeurer
1df5fed50a [turbofan] Cleanup the NodeProperties.
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/883613006

Cr-Commit-Position: refs/heads/master@{#26316}
2015-01-29 09:18:09 +00:00
jarin
5bd8407f8c Reland of "Steps towards unification of number bitset and range types."
This reverts commit 7619374979.

BUG=

Review URL: https://codereview.chromium.org/877643002

Cr-Commit-Position: refs/heads/master@{#26301}
2015-01-28 08:42:32 +00:00
danno
c7b09aac31 Remove the dependency of Zone on Isolate
Along the way:
- Thread isolate parameter explicitly through code that used to
  rely on getting it from the zone.
- Canonicalize the parameter position of isolate and zone for
  affected code
- Change Hydrogen New<> instruction templates to automatically
  pass isolate

R=mstarzinger@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/868883002

Cr-Commit-Position: refs/heads/master@{#26252}
2015-01-23 15:20:00 +00:00
bmeurer
7619374979 Revert of Steps towards unification of number bitset and range types. (patchset #4 id:60001 of https://codereview.chromium.org/837723006/)
Reason for revert:
Breaks test-types/Maybe, i.e.

out/Release/cctest --random-seed=-707413401 test-types/Maybe

started failing afterwards

Original issue's description:
> Steps towards unification of number bitset and range types.
>
> - New invariant on union types: if the union has a range then the number
>   bits in the bitset must be cleared.
>
> - Various tweaks in intersection and union to satisfy the invariant.
>
> - Exposed and used representation bits in range types (and the Limits
>   helper class).
>
> - Implemented Glb for ranges so that the Is predicate handles
>   ranges correctly.
>
> - Change typer weakening so that it does not rely on GetRange.
>   However, the code still seems to be a bit fragile.
>
> - Removed the Smi types from the type system core, instead introduced
>   Signed31, Unsigned30 and created constructors for Small(Un)Signed
>   that point to the right type for the architecture.
>
> - Punched a hole in the config to be able to get to the isolate so
>   that it is possible to allocate heap numbers for newly created
>   ranges.
>
> Patch by jarin@chromium.prg, original review here:
> https://codereview.chromium.org/795713003/
>
> TBR=jarin@chromium.org
> BUG=
>
> Committed: https://crrev.com/2764fd8d1a266a9136c987c2483492113b0c8d80
> Cr-Commit-Position: refs/heads/master@{#26197}

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

Review URL: https://codereview.chromium.org/868583002

Cr-Commit-Position: refs/heads/master@{#26207}
2015-01-22 05:33:39 +00:00
rossberg
2764fd8d1a Steps towards unification of number bitset and range types.
- New invariant on union types: if the union has a range then the number
  bits in the bitset must be cleared.

- Various tweaks in intersection and union to satisfy the invariant.

- Exposed and used representation bits in range types (and the Limits
  helper class).

- Implemented Glb for ranges so that the Is predicate handles
  ranges correctly.

- Change typer weakening so that it does not rely on GetRange.
  However, the code still seems to be a bit fragile.

- Removed the Smi types from the type system core, instead introduced
  Signed31, Unsigned30 and created constructors for Small(Un)Signed
  that point to the right type for the architecture.

- Punched a hole in the config to be able to get to the isolate so
  that it is possible to allocate heap numbers for newly created
  ranges.

Patch by jarin@chromium.prg, original review here:
https://codereview.chromium.org/795713003/

TBR=jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/837723006

Cr-Commit-Position: refs/heads/master@{#26197}
2015-01-21 15:29:22 +00:00
bmeurer
c329a49d69 [turbofan] Fix incorrect minus zero handling in the unit tests.
TEST=unittests
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/831203003

Cr-Commit-Position: refs/heads/master@{#25956}
2015-01-05 12:29:27 +00:00
bmeurer
00013a5692 [turbofan] Minor cleanup to reduce code duplication.
TEST=unittests
R=jochen@chromium.org

Review URL: https://codereview.chromium.org/819103002

Cr-Commit-Position: refs/heads/master@{#25927}
2014-12-22 14:37:22 +00:00
machenbach
7d478d9621 Revert of Remove obsolete V8_INFINITY macro. (patchset #3 id:40001 of https://codereview.chromium.org/798413003/)
Reason for revert:
Speculative revert. This seems to block the current roll: https://codereview.chromium.org/819653003/

I retried several times, also with a new roll. The error is internal - but that doesn't make much of a difference.

Original issue's description:
> Remove obsolete V8_INFINITY macro.
>
> Use std::numeric_limits consistently.
>
> R=svenpanne@chromium.org
>
> Committed: https://crrev.com/31c66e2d53569c4e229d55483d28208491e73612
> Cr-Commit-Position: refs/heads/master@{#25897}

TBR=svenpanne@chromium.org,bmeurer@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/813813003

Cr-Commit-Position: refs/heads/master@{#25912}
2014-12-20 13:17:35 +00:00
bmeurer
31c66e2d53 Remove obsolete V8_INFINITY macro.
Use std::numeric_limits consistently.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/798413003

Cr-Commit-Position: refs/heads/master@{#25897}
2014-12-19 07:18:00 +00:00
jarin
5913c7cc3f Reland of "Avoid number range holes in bitset types."
This reverts commit 8a6cbf0a86.

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/788313002

Cr-Commit-Position: refs/heads/master@{#25786}
2014-12-11 17:50:11 +00:00
jarin
8a6cbf0a86 Revert of Avoid number range holes in bitset types. (patchset #5 id:80001 of https://codereview.chromium.org/759013003/)
Reason for revert:
For breaking the waterfall (run-json-stringify test).

Original issue's description:
> Avoid number range holes in bitset types.
>
> BUG=

TBR=rossberg@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/794663002

Cr-Commit-Position: refs/heads/master@{#25756}
2014-12-10 18:25:38 +00:00
jarin
4de4f74cf2 Avoid number range holes in bitset types.
BUG=

Review URL: https://codereview.chromium.org/759013003

Cr-Commit-Position: refs/heads/master@{#25754}
2014-12-10 16:39:10 +00:00
titzer@chromium.org
b9886ae9ff Fix bugs in simplified lowering relating to int32/uint32 signs.
Lowering of NumberToUint32 and NumberToInt32 was not correctly accounting for the sign of the input and the sign of the output, emitting the wrong representation changes.

Along the way, I've found cases where MachineOperatorBuilder would break if fed a machine type for loads or stores that was not cached, requiring MachineOperatorBuilder to take zone to allocate operators for these cases.

R=bmeurer@chromium.org, jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/714613002

Cr-Commit-Position: refs/heads/master@{#25247}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25247 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-10 14:28:42 +00:00
bmeurer@chromium.org
8798c410e1 [turbofan] Turn various diamonds into selects.
TEST=cctest/test-changes-lowering,mjsunit/asm,unittests
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/704463004

Cr-Commit-Position: refs/heads/master@{#25180}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25180 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-06 09:10:32 +00:00
sigurds@chromium.org
75ac43928b Add floor, ceil, round (truncate) instructions for ia32, x64 (if SSE4.1) and
add floor, ceil, round (truncate and away from zero) for arm64.

R=bmeurer@chromium.org, dcarney@chromium.org, mstarzinger@chromium.org, rodolph.perfetta@arm.com
TEST=test/mjsunit/asm/math-floor.js,test/mjsunit/asm/math-ceil.js,test/unittest/compiler/js-builtin-reducer-unittest.cc

Review URL: https://codereview.chromium.org/677433002

Cr-Commit-Position: refs/heads/master@{#25018}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-30 14:16:15 +00:00
bmeurer@chromium.org
cc60a45d78 [turbofan] Move node matchers to separate file.
TEST=unittests
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/639293006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24729 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-20 11:26:23 +00:00
rossberg@chromium.org
0e16150d33 Better typing and type verification
- Extend verifier to check types of JS and Simplified nodes.
- Untyped nodes now contain NULL as types, enforcing hard failure.
- Typer immediately installs itself as a decorator; remove explicit decorator installation.
- Decorator eagerly types all nodes that have typed inputs
  (subsumes typing of constant cache, removing its typing
  side-channel and various spurious dependencies on the typer).
- Cut down typer interface to prevent inconsistently typed graphs.
- Remove verification from start, since it caused too much trouble
  with semi-wellformed nodes.
- Fix a couple of bugs on the way that got uncovered.

To do: verifying machine operators. Also, various conditions in the
verifier are currently commented out, because they don't yet hold.

BUG=
R=jarin@chromium.org,titzer@chromium.org

Review URL: https://codereview.chromium.org/658543002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24626 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-15 11:38:04 +00:00
mstarzinger@chromium.org
f99fd3867b Remove premordial math functions from native context.
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/635323002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24471 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-08 14:42:31 +00:00
bmeurer@chromium.org
e3294b1f09 [turbofan] Fix lowering of typed loads/stores.
Only JSLoadProperty/JSStoreProperty nodes with external typed arrays can
be lowered to LoadElement/StoreElement, because lowering of non-external
typed arrays would require a map check.

TEST=cctest,unittests,mjsunit
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/631093003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24426 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-07 07:36:21 +00:00
jarin@chromium.org
f40d582cf1 Revert "[turbofan] Fix lowering of typed loads/stores."
This reverts commit r24386 for tanking asm.js benchmarks.

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/634473002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24406 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-06 08:54:24 +00:00
bmeurer@chromium.org
5899cc8ca7 [turbofan] Fix lowering of typed loads/stores.
We can only access to external typed arrays; lowering of internal
typed arrays would require a map check plus eager deoptimization.
Also embed the array buffer reference directly instead of embedding
the typed array.

TEST=cctest,mjsunit,unittests
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/621863002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24386 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-02 08:38:37 +00:00
bmeurer@chromium.org
bfd37ab267 Move unit tests to test/unittests.
As per discussion on the V8 team, this is the place we want them to live,
not following the Chrome Style Guide for this.

BUG=v8:3489
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/615393002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24350 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-01 08:34:25 +00:00