Commit Graph

257 Commits

Author SHA1 Message Date
fschneider@google.com f7a97fa4d4 VM: Fix slow-path deoptimization environment.
The number of pushed arguments needs to be accounted for.

BUG=dartbug.com/22331
R=vegorov@google.com

Review URL: https://codereview.chromium.org//914133002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43681 260f80e4-7a28-3924-810f-c04153c831b5
2015-02-11 13:34:15 +00:00
srdjan@google.com 81d8099650 Redesign inlining interval computation.
Review URL: https://codereview.chromium.org//904763003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43622 260f80e4-7a28-3924-810f-c04153c831b5
2015-02-09 23:11:05 +00:00
johnmccutchan@google.com 635ba4bbd9 Port flow_graph*, disassembler to ISL_Print
BUG=
R=srdjan@google.com

Review URL: https://codereview.chromium.org//907093002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43609 260f80e4-7a28-3924-810f-c04153c831b5
2015-02-09 20:49:32 +00:00
zra@google.com 1299ad1485 Allows turning on checked mode on a per-isolate basis
using the embedding API.

R=regis@google.com, srdjan@google.com

Review URL: https://codereview.chromium.org//883263004

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43481 260f80e4-7a28-3924-810f-c04153c831b5
2015-02-04 18:32:48 +00:00
fschneider@google.com 75ad2d8253 VM: Improve polymorphic check-class in optimized code.
Sort comparisons by decreasing IC count like we do for polymorphic inlining.

R=srdjan@google.com

Review URL: https://codereview.chromium.org//895603002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43392 260f80e4-7a28-3924-810f-c04153c831b5
2015-02-03 12:30:51 +00:00
srdjan@google.com 12c5e2bbaf Cleanup: use const reference for ParsedFunction where possible.
R=rmacnak@google.com

Review URL: https://codereview.chromium.org//878243002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43226 260f80e4-7a28-3924-810f-c04153c831b5
2015-01-28 00:41:09 +00:00
srdjan@google.com 110a74f954 Cleanups: parsed_function()->function() => function()
R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org//881063003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@43213 260f80e4-7a28-3924-810f-c04153c831b5
2015-01-27 22:09:31 +00:00
koda@google.com 33b55e64e4 Isolate/Thread split: Isolate -> Zone for LocationSummary.
The isolate was being passed around, even though only its current zone was used.

Pass that zone directly instead, giving two benefits:
1. helps prepare for the upcoming Isolate/Thread split, where "new(isolate)" must go, and
2. saves a pointer indirection at the allocation sites.

Review URL: https://codereview.chromium.org//855533002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42935 260f80e4-7a28-3924-810f-c04153c831b5
2015-01-15 21:29:57 +00:00
zerny@google.com 5c04899198 Define a lower threshold for optimizing irregexp generated functions.
The threshold is given by the flag --regexp_optimization_count_threshold
and has a default value of 1000.

R=fschneider@google.com
BUG=

Review URL: https://codereview.chromium.org//836703003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42632 260f80e4-7a28-3924-810f-c04153c831b5
2015-01-06 13:56:21 +00:00
srdjan@google.com 4c8545557a Add inlining ranges/intervals to code objects so that we can map a pc to the inlined stack. The mapping is not fast, used only for disassembly.
Copied from other CL: https://codereview.chromium.org/790213004/

Review URL: https://codereview.chromium.org//817823002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42522 260f80e4-7a28-3924-810f-c04153c831b5
2014-12-19 20:03:27 +00:00
vegorov@google.com 9fcfaa8468 Support remaping of pair location inside materializations for slow paths.
Move remaping logic into Location::RemapForSlowPath to ensure that it is shared and will not get out of sync again.

BUG=http://dartbug.com/21887
R=fschneider@google.com

Review URL: https://codereview.chromium.org//811823002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42396 260f80e4-7a28-3924-810f-c04153c831b5
2014-12-16 15:18:51 +00:00
fschneider@google.com 9fad07f46e Enable inlining inside try-blocks in the VM's optimizing compiler.
R=vegorov@google.com

Review URL: https://codereview.chromium.org//773183002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42103 260f80e4-7a28-3924-810f-c04153c831b5
2014-12-04 12:29:28 +00:00
srdjan@google.com bad6a61899 Cleanups.
R=vegorov@google.com

Review URL: https://codereview.chromium.org//774763002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@42069 260f80e4-7a28-3924-810f-c04153c831b5
2014-12-02 18:48:12 +00:00
zerny@google.com 8e807c8550 Integrate the Irregexp Regular Expression Engine.
BUG=http://dartbug.com/19090
R=fschneider@google.com

Committed: https://code.google.com/p/dart/source/detail?r=41949

Review URL: https://codereview.chromium.org//744853003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41983 260f80e4-7a28-3924-810f-c04153c831b5
2014-11-26 09:32:43 +00:00
zerny@google.com 84a4135901 Revert "Integrate the Irregexp Regular Expression Engine."
This reverts commit https://code.google.com/p/dart/source/detail?r=41949

TBR=fschneider@google.com
BUG=

Review URL: https://codereview.chromium.org//754383002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41950 260f80e4-7a28-3924-810f-c04153c831b5
2014-11-25 13:36:40 +00:00
zerny@google.com 43ac0c6f33 Integrate the Irregexp Regular Expression Engine.
BUG=http://dartbug.com/19090
R=fschneider@google.com

Review URL: https://codereview.chromium.org//744853003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41949 260f80e4-7a28-3924-810f-c04153c831b5
2014-11-25 13:10:21 +00:00
vegorov@google.com e6847ac17c IR refactoring: consolidate all boxing and unboxing instructions.
BUG=
R=fschneider@google.com

Review URL: https://codereview.chromium.org//670263007

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41437 260f80e4-7a28-3924-810f-c04153c831b5
2014-10-31 12:17:12 +00:00
fschneider@google.com 876193b31a Make CTX allocatable by the register allocator.
This change makes CTX available by not caching the current
context while in Dart code. Instead the current context
is held in a local variable (:saved_current_context_var) and
is passed as argument in CTX at calls.

This also simplifies a lot of code in the debugger: As a result,
Isolate::top_context is not needed anymore since the current context
can always be extracted from a Dart frame.

R=vegorov@google.com

Review URL: https://codereview.chromium.org//678763004

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41422 260f80e4-7a28-3924-810f-c04153c831b5
2014-10-30 15:42:38 +00:00
vegorov@google.com dde7963d36 Ignore redundant parallel moves when checking if a block is empty.
Single predecessor blocks containing nothing but redundant parallel moves can be compacted away.

R=zerny@google.com
BUG=http://dartbug.com/21302

Review URL: https://codereview.chromium.org//669613003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41187 260f80e4-7a28-3924-810f-c04153c831b5
2014-10-20 11:29:19 +00:00
fschneider@google.com e6ef4b16bf Fix captured parameters and optimized try-catch.
The stack slots of captured parameters must be skipped when
generating sync code in optimized try-catch. Since those parameters
are initially copied into the context, their values are not recorded
in the environment.

Store-to-load forwarding may though use the original initial value
from the stack and therefore it must not be overwritten by try-sync
code that predeeds every call inside optimized try-blocks.

R=srdjan@google.com

Review URL: https://codereview.chromium.org//653073002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@41099 260f80e4-7a28-3924-810f-c04153c831b5
2014-10-14 14:48:32 +00:00
vegorov@google.com e6113cc0f5 If bounds check 0 <= index < length is not redundant we attempt to
replace it with a sequence of checks that guarantee

          0 <= LowerBound(index) < UpperBound(index) < length

and hoist all of those checks out of the enclosing loop.

Upper/Lower bounds are symbolic arithmetic expressions with +, -, *
operations and are computed based on discovered simple induction variables.

Simple induction variable is a variable that follows the pattern v1 <- phi(v0, v1 + 1)

BUG=
R=fschneider@google.com

Review URL: https://codereview.chromium.org//619903002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40969 260f80e4-7a28-3924-810f-c04153c831b5
2014-10-07 18:50:29 +00:00
koda@google.com 60b12996ec Track references to allocation stubs via static_calls_table, instead of keeping two referencers alive.
(Verbatim copy of srdjan's https://codereview.chromium.org/609893002/ )

TBR=srdjan@google.com

Review URL: https://codereview.chromium.org//608913002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40738 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-26 20:17:53 +00:00
fschneider@google.com bdd6b29fd8 Revert r40306 because of crash.
TBR=vegorov@google.com

Review URL: https://codereview.chromium.org//572913002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40312 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-16 12:14:33 +00:00
fschneider@google.com cd78dbe709 Refactor generating lazy deoptimization descriptors.
Instead of having special handling for closure calls, pass
the number of input operands down so that the correct lazy
deoptimization environment is generated at calls.

There is no functional change here - this is to simplify
having calls with pushed arguments and input operands
(like closure calls).

Runtime calls still do not have a correct lazy deoptimization environment
the environment should only be used for creating stack traces
from optimized code.

Fix a bug with --trace-deoptimization-verbose printing: We
cannot call ToCString on objects when the slots have not
been filled in yet. (e.g. Closures)

R=vegorov@google.com

Review URL: https://codereview.chromium.org//575443002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40306 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-16 10:31:47 +00:00
fschneider@google.com 2731828467 Refactor emitting of deoptimization descriptors for calls.
Instead of emitting the descriptor in EmitNativeCode, the
instruction prologue now emits the descriptor automatically.
To do that all instructions that need such a descriptor have
CanBecomeDeoptimizationTarget set to true.

I moved the duplicated for TargetEntryInstr::EmitNativeCode
to intermediate_language.cc.

R=vegorov@google.com

Review URL: https://codereview.chromium.org//565643002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40150 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-11 13:54:20 +00:00
fschneider@google.com da0fb6bbe2 More intrinsics in IR.
R=srdjan@google.com, vegorov@google.com, zra@google.com

Review URL: https://codereview.chromium.org//513213002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@40088 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-10 13:34:58 +00:00
fschneider@google.com 0469c12072 VM: Remove unused IsClosureCall bit from IC data.
Since closure calls are now implemented in terms of instance calls
 (via .call and a dispatcher method), there is no need for a bit in
  the IC data tracking if a IC call may be a closure call.

R=vegorov@google.com

Review URL: https://codereview.chromium.org//551043002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39979 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-08 17:10:59 +00:00
fschneider@google.com 6b4c05766f Fix xcode build for real.
R=vegorov@google.com

Review URL: https://codereview.chromium.org//540303002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39909 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-05 11:30:38 +00:00
fschneider@google.com 36c0b7b0e8 Attempt to fix xcode build.
tbr=vegorov@google.com

Review URL: https://codereview.chromium.org//546673003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39906 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-05 10:48:17 +00:00
fschneider@google.com 08b9e1fb2d Fixes for ARM/ARM64/MIPS parallel move resolver.
The parallel move resolve has to block reserved registers like
TMP, TMP2, PP, CTX so that they are not used as scratch registers.

Use allocated scratch register instead of IP because IP may be
used for large stack slot offsets.

This CL also generalizes stack-slot locations to address
stack lots relative to FP or SP.

R=zra@google.com

Review URL: https://codereview.chromium.org//539073002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39905 260f80e4-7a28-3924-810f-c04153c831b5
2014-09-05 09:49:53 +00:00
ajohnsen@google.com ff1b4190c1 Scale invocation count by the number of BBs in the flow-graph of a method.
BUG=
R=vegorov@google.com

Review URL: https://codereview.chromium.org//501553005

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39698 260f80e4-7a28-3924-810f-c04153c831b5
2014-08-29 12:51:09 +00:00
srdjan@google.com 2a3c2ad058 Add loop printing to block comments.
R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org//515913003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39673 260f80e4-7a28-3924-810f-c04153c831b5
2014-08-28 19:52:47 +00:00
srdjan@google.com 85af0719ef Faster IC stubs by specializing them for two-argument Smi operations (implemented only a few for now)
R=zra@google.com

Review URL: https://codereview.chromium.org//436643002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39471 260f80e4-7a28-3924-810f-c04153c831b5
2014-08-21 20:08:40 +00:00
fschneider@google.com 36c848d1d7 VM: Optimized context allocation on all platforms.
This change generate inline code for context allocation and makes
the stores to initialize the parent pointer and the context variables
explicit in the optimized flow graph. This allows dead store elimination
to eliminate those stores.

Before, we only had inlined context allocation for ia32.

Furthermore:
* Added an assertion that the exit label of slow-path code
is bound when emitting deferred code. This used to cause random crashes
if forgotten.

* Fixed a bug in allocation stats on ARM.

BUG=dartbug.com/17238
R=srdjan@google.com

Review URL: https://codereview.chromium.org//346823003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@39202 260f80e4-7a28-3924-810f-c04153c831b5
2014-08-13 15:49:01 +00:00
srdjan@google.com 00891583b0 Minor cleanup: remove unnecessary argument.
R=regis@google.com

Review URL: https://codereview.chromium.org//422893003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@38758 260f80e4-7a28-3924-810f-c04153c831b5
2014-07-30 22:50:38 +00:00
vegorov@google.com 36cddbd749 Use shorter TryAllocate instruction sequence on ARM/ARM64/MIPS.
On these architectures it's preferable to load Scavenger address into the register and then access top/end fields through this register instead of loading addresses of those fields as immediates --- which takes either two instructions or a memory load through a pool pointer.

Cleanup boxing slow-paths in the optimizing compiler. We had tons of duplicated code that was essentially doing the same thing.

BUG=
R=johnmccutchan@google.com, regis@google.com

Review URL: https://codereview.chromium.org//410333003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@38539 260f80e4-7a28-3924-810f-c04153c831b5
2014-07-24 12:04:21 +00:00
koda@google.com aca839c54f Allocate fewer empty arrays.
Use VM isolate's canonical, untyped, empty array when possible.

Takes Dart2JS benchmark from ~9000 empty arrays to ~200.

R=asiva@google.com

Review URL: https://codereview.chromium.org//381483002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@38435 260f80e4-7a28-3924-810f-c04153c831b5
2014-07-21 16:41:05 +00:00
asiva@google.com f85f986bda Make isolate specific stub code accessors instance methods instead
of static. This allows use of isolate without having to call
Isolate::Current() when using these accessors.

R=johnmccutchan@google.com, srdjan@google.com, zra@google.com

Review URL: https://codereview.chromium.org//365983002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@38040 260f80e4-7a28-3924-810f-c04153c831b5
2014-07-07 21:25:00 +00:00
srdjan@google.com 2bb77cd8d5 Iterate over PcDescriptors only via iterators, not via an index. (preparation for more compression of PcDescriptors).
R=asiva@google.com

Review URL: https://codereview.chromium.org//356923006

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@38032 260f80e4-7a28-3924-810f-c04153c831b5
2014-07-07 18:25:45 +00:00
regis@google.com 32f3e11d99 Cleanup of error and warning reporting.
R=hausner@google.com, srdjan@google.com

Review URL: https://codereview.chromium.org//340203003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37468 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-18 22:30:34 +00:00
srdjan@google.com 8f2dc961d0 Save ICData of unoptimized code in the function, thus preserving it across repated unoptimized compilations.
R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org//335173002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37378 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-16 22:24:05 +00:00
ngeoffray@google.com 4936f84613 Fix a bug in ParallelMoveResolver::AllocateScratchRegister, where we end up allocating PC on ARM.
1) Iterate from first_free_register to last_free_register to ensure we're not using protected registers.
2) Break when finding an ok register.

2) fixes allocating PC on ARM by itself, but we could need 1) for another architecture.

The reason this has not been exposed yet, is because we call this method for a memory <-> memory
swap, which almost never happens on ARM.

R=vegorov@google.com

Review URL: https://codereview.chromium.org//337503003

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37304 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-13 08:02:23 +00:00
fschneider@google.com 37d2d57532 VM: Reduce the number of edge counters in generated code.
This CL avoids edge counters for the incoming edge of blocks that
 have one predecessor and one succesor block. The incoming count is
 always equal to the outgoing count (modulo exceptional exit by throwing).

R=johnmccutchan@google.com

Review URL: https://codereview.chromium.org//311993007

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37170 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-10 10:27:45 +00:00
regis@google.com 3617cf8077 Share ic data between unoptimized and optimized static calls.
This prevents duplicated javascript compatibility warnings.

R=srdjan@google.com

Review URL: https://codereview.chromium.org//322633002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37103 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-06 21:47:51 +00:00
regis@google.com 139187ae5f Add javascript compatibility warnings for strings that are not identical, but
equal. The vm does not canonicalize all strings as javascript does.

R=srdjan@google.com

Review URL: https://codereview.chromium.org//312693008

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@37013 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-04 20:47:54 +00:00
fschneider@google.com 559e6410e2 First step in reducing the size of PC descriptors.
The descriptors used for patching just contain a pc offset
and can be stored with the code object instead of allocating
a full descriptor and linearly searching the list to retrieve
this information.

R=srdjan@google.com

Review URL: https://codereview.chromium.org//315583002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@36989 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-04 11:02:34 +00:00
vegorov@google.com 4e286b316f Split GuardField into GuardFieldType and GuardFieldLength instructions.
Record expected length offset that matches guarded_cid.

This greatly simplifies code generation and control flow inside these guards.

Fixes the bug that was causing unoptimized code to call into runtime even when guard did not fail.

Add --trace-field-guards flags to trace updates to the state of guarded fields' properties.

BUG=http://dartbug.com/19003
R=fschneider@google.com, johnmccutchan@google.com, zra@google.com

Review URL: https://codereview.chromium.org//304703002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@36872 260f80e4-7a28-3924-810f-c04153c831b5
2014-06-02 13:53:53 +00:00
zra@google.com 6143dd9bad Adds more ARM64 SIMD instructions.
Also, enables simd inlining for arm64.

R=regis@google.com

Review URL: https://codereview.chromium.org//307523002

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@36694 260f80e4-7a28-3924-810f-c04153c831b5
2014-05-27 16:29:15 +00:00
iposva@google.com d36cd610cb - Use isolate where it is appropriate.
- Do not emit an name for ExternalLabel.

R=johnmccutchan@google.com, srdjan@google.com

Review URL: https://codereview.chromium.org//306483005

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@36662 260f80e4-7a28-3924-810f-c04153c831b5
2014-05-27 10:15:50 +00:00
srdjan@google.com 45692c1248 More reduction of accessing TLS by caching/passing isolate value.
R=regis@google.com

Review URL: https://codereview.chromium.org//299263005

git-svn-id: https://dart.googlecode.com/svn/branches/bleeding_edge/dart@36598 260f80e4-7a28-3924-810f-c04153c831b5
2014-05-23 23:26:48 +00:00