VM: Optimized context allocation on all platforms.

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

Committed: https://code.google.com/p/dart/source/detail?r=39202

[Florian Schneider, 2014-08-06 10:52:42]

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/intermediate_...
File runtime/vm/intermediate_language_arm.cc (left):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/intermediate_...
runtime/vm/intermediate_language_arm.cc:2227: __ LoadImmediate(R8,
heap->TopAddress());
This is fixed in TryAllocateArray (assembler_arm.cc).
It should be 
__ LoadImmediate(R8, kArraySize);

[Florian Schneider, 2014-08-11 09:06:04]

On 2014/08/06 10:52:42, Florian Schneider wrote:
>
https://codereview.chromium.org/346823003/diff/60001/runtime/vm/intermediate_...
> File runtime/vm/intermediate_language_arm.cc (left):
> 
>
https://codereview.chromium.org/346823003/diff/60001/runtime/vm/intermediate_...
> runtime/vm/intermediate_language_arm.cc:2227: __ LoadImmediate(R8,
> heap->TopAddress());
> This is fixed in TryAllocateArray (assembler_arm.cc).
> It should be 
> __ LoadImmediate(R8, kArraySize);

Ping.

[srdjan, 2014-08-11 21:42:34]

LGTM

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm.cc
File runtime/vm/assembler_arm.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
runtime/vm/assembler_arm.cc:3190: Register temp2) {
Guard with FLAG_inline_alloc

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
File runtime/vm/assembler_arm64.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
runtime/vm/assembler_arm64.cc:1423: Isolate* isolate = Isolate::Current();
Guard this with FLAG_inline_alloc.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_mip...
File runtime/vm/assembler_mips.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_mip...
runtime/vm/assembler_mips.cc:896: Register temp2) {
Guard with FLAG_inline_alloc.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_x64.cc
File runtime/vm/assembler_x64.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_x64...
runtime/vm/assembler_x64.cc:3137:
ClassHeapStats::allocated_size_since_gc_new_space_offset() :
This and above and below should have 4 spaces indentation.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/flow_graph_op...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/flow_graph_op...
runtime/vm/flow_graph_optimizer.cc:4413: void
FlowGraphOptimizer::VisitAllocateContext(AllocateContextInstr* instr) {
Why don't you do that in the FlowGraphBuilder instead?

[Florian Schneider, 2014-08-13 14:56:12]

Addressed comment. I also made sure that initializing stores are safely handled
by dead code elimination.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm.cc
File runtime/vm/assembler_arm.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
runtime/vm/assembler_arm.cc:3190: Register temp2) {
On 2014/08/11 21:42:34, srdjan wrote:
> Guard with FLAG_inline_alloc

Done.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
File runtime/vm/assembler_arm64.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_arm...
runtime/vm/assembler_arm64.cc:1423: Isolate* isolate = Isolate::Current();
On 2014/08/11 21:42:34, srdjan wrote:
> Guard this with FLAG_inline_alloc.

Done.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_mip...
File runtime/vm/assembler_mips.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_mip...
runtime/vm/assembler_mips.cc:896: Register temp2) {
On 2014/08/11 21:42:34, srdjan wrote:
> Guard with FLAG_inline_alloc.

Done.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_x64.cc
File runtime/vm/assembler_x64.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_x64...
runtime/vm/assembler_x64.cc:3137:
ClassHeapStats::allocated_size_since_gc_new_space_offset() :
On 2014/08/11 21:42:34, srdjan wrote:
> This and above and below should have 4 spaces indentation.

Done.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/assembler_x64...
runtime/vm/assembler_x64.cc:3137:
ClassHeapStats::allocated_size_since_gc_new_space_offset() :
On 2014/08/11 21:42:34, srdjan wrote:
> This and above and below should have 4 spaces indentation.

Done.

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/flow_graph_op...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/346823003/diff/60001/runtime/vm/flow_graph_op...
runtime/vm/flow_graph_optimizer.cc:4413: void
FlowGraphOptimizer::VisitAllocateContext(AllocateContextInstr* instr) {
On 2014/08/11 21:42:34, srdjan wrote:
> Why don't you do that in the FlowGraphBuilder instead?

I only want the inlined code for unoptimized code for code size reasons.
Performance-wise it only matters in optimized code anyway.

[Florian Schneider, 2014-08-13 15:49:19]

Committed patchset #5 manually as r39202 (presubmit successful).