[turbofan]: Integrate basic type feedback for property accesses.

src/flag-definitions.h

 // Copyright 2012 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 // This file defines all of the flags.  It is separated into different section,
 // for Debug, Release, Logging and Profiling, etc.  To add a new flag, find the
 // correct section, and use one of the DEFINE_ macros, without a trailing ';'.
 //
 // This include does not have a guard, because it is a template-style include,
 // which can be included multiple times in different modes.  It expects to have
@@ skipping 382 matching lines @@
               "trace turbo cfg graph (for C1 visualizer) to a given file name")
 DEFINE_BOOL(trace_turbo_types, true, "trace TurboFan's types")
 DEFINE_BOOL(trace_turbo_scheduler, false, "trace TurboFan's scheduler")
 DEFINE_BOOL(trace_turbo_reduction, false, "trace TurboFan's various reducers")
 DEFINE_BOOL(trace_turbo_jt, false, "trace TurboFan's jump threading")
 DEFINE_BOOL(turbo_asm, true, "enable TurboFan for asm.js code")
 DEFINE_BOOL(turbo_verify, DEBUG_BOOL, "verify TurboFan graphs at each phase")
 DEFINE_BOOL(turbo_stats, false, "print TurboFan statistics")
 DEFINE_BOOL(turbo_splitting, true, "split nodes during scheduling in TurboFan")
 DEFINE_BOOL(turbo_types, true, "use typed lowering in TurboFan")
+DEFINE_BOOL(turbo_type_feedback, false, "use type feedback in TurboFan")
 DEFINE_BOOL(turbo_source_positions, false,
             "track source code positions when building TurboFan IR")
 DEFINE_IMPLICATION(trace_turbo, turbo_source_positions)
+DEFINE_IMPLICATION(turbo_type_feedback, turbo_deoptimization)

[titzer, 2015/03/24 10:49:32]

Comments on this? Should type feedback imply deoptimization support? We could
still support fast path map checks and guarded inlining without deoptimizing,
but nothing that requires code dependencies.

[Michael Starzinger, 2015/03/24 12:11:18]

Agreed, it would be nice to be able to leverage dynamic feedback even without
deoptimization support. Not sure if we ever want to run in such a mode, but it
would be nice to know that we could. :)

[titzer, 2015/03/24 14:52:06]

I removed the implication and put a branch in js-type-feedback.cc

 DEFINE_BOOL(context_specialization, false,
             "enable context specialization in TurboFan")
 DEFINE_BOOL(turbo_deoptimization, false, "enable deoptimization in TurboFan")
 DEFINE_BOOL(turbo_inlining, false, "enable inlining in TurboFan")
 DEFINE_BOOL(turbo_builtin_inlining, true, "enable builtin inlining in TurboFan")
 DEFINE_BOOL(trace_turbo_inlining, false, "trace TurboFan inlining")
 DEFINE_BOOL(loop_assignment_analysis, true, "perform loop assignment analysis")
 DEFINE_BOOL(turbo_profiling, false, "enable profiling in TurboFan")
 // TODO(dcarney): this is just for experimentation, remove when default.
 DEFINE_BOOL(turbo_delay_ssa_decon, false,
@@ skipping 619 matching lines @@
 #undef DEFINE_ALIAS_FLOAT
 #undef DEFINE_ALIAS_ARGS
 
 #undef FLAG_MODE_DECLARE
 #undef FLAG_MODE_DEFINE
 #undef FLAG_MODE_DEFINE_DEFAULTS
 #undef FLAG_MODE_META
 #undef FLAG_MODE_DEFINE_IMPLICATIONS
 
 #undef COMMA