Add TraceBuffer log to all Functions (plus small tweaks to function-view)

Add TraceBuffer log to all Functions (plus small tweaks to function-view)

BUG=
R=srdjan@google.com, turnidge@google.com

Committed: https://code.google.com/p/dart/source/detail?r=35207

[srdjan, 2014-04-14 15:02:52]

LGTM.

[turnidge, 2014-04-14 16:41:16]

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
File runtime/vm/deopt_instructions.cc (right):

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
runtime/vm/deopt_instructions.cc:346: const Code& code =
Code::Handle(top_frame->LookupDartCode());
Instead of using a DartFrameIterator to get the code here, the DeoptContext has
a code() accessor -- can you see if that works instead?

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
runtime/vm/deopt_instructions.cc:350: DeoptReasonToText(deopt_reason()));
We should eventually include which code object was deoptimized in the log --
maybe add a TODO?

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc#newcode...
runtime/vm/object.cc:5893: result.set_log(new TraceBuffer());
Each function gets its own TraceBuffer and the default capacity of a TraceBuffer
is 1024 entries.  That seems like a lot given that each function will probably
have less than 10 messages.

If we are going to have many TraceBuffers, we might want -- for example -- to
start them with a small capacity and double the capacity on demand up to some
max.

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc#newcode...
runtime/vm/object.cc:6461: jsobj.AddProperty("optimized", false);
What do you think of dropping "optimized" -- is it enough to check for whether
we have and "optimizedCode"?

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc#newcode...
runtime/vm/object.cc:6462: }
Should we add OSR code objects too?

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.h#newcode1901
runtime/vm/object.h:1901: }
I've been pondering how we can attach json objects to the logs.  We should
discuss.

[Cutch, 2014-04-15 20:47:57]

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
File runtime/vm/deopt_instructions.cc (right):

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
runtime/vm/deopt_instructions.cc:346: const Code& code =
Code::Handle(top_frame->LookupDartCode());
On 2014/04/14 16:41:17, turnidge wrote:
> Instead of using a DartFrameIterator to get the code here, the DeoptContext
has
> a code() accessor -- can you see if that works instead?

Done.

https://codereview.chromium.org/227423005/diff/1/runtime/vm/deopt_instruction...
runtime/vm/deopt_instructions.cc:350: DeoptReasonToText(deopt_reason()));
On 2014/04/14 16:41:17, turnidge wrote:
> We should eventually include which code object was deoptimized in the log --
> maybe add a TODO?

Done.

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc#newcode...
runtime/vm/object.cc:6461: jsobj.AddProperty("optimized", false);
On 2014/04/14 16:41:17, turnidge wrote:
> What do you think of dropping "optimized" -- is it enough to check for whether
> we have and "optimizedCode"?

Done.

https://codereview.chromium.org/227423005/diff/1/runtime/vm/object.cc#newcode...
runtime/vm/object.cc:6462: }
On 2014/04/14 16:41:17, turnidge wrote:
> Should we add OSR code objects too?

Discussed offline.

[turnidge, 2014-04-15 21:35:20]

lgtm w/ optional comments below.

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer.cc
File runtime/vm/trace_buffer.cc (right):

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer....
runtime/vm/trace_buffer.cc:13: ring_max_capacity_(maximum_capacity) {
ring_capacity -> capacity, etc.

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer....
runtime/vm/trace_buffer.cc:33: void TraceBuffer::Resize(intptr_t capacity) {
capacity -> new_capacity?

[Cutch, 2014-04-21 15:39:19]

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer.cc
File runtime/vm/trace_buffer.cc (right):

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer....
runtime/vm/trace_buffer.cc:13: ring_max_capacity_(maximum_capacity) {
On 2014/04/15 21:35:20, turnidge wrote:
> ring_capacity -> capacity, etc.

Done.

https://codereview.chromium.org/227423005/diff/10001/runtime/vm/trace_buffer....
runtime/vm/trace_buffer.cc:33: void TraceBuffer::Resize(intptr_t capacity) {
On 2014/04/15 21:35:20, turnidge wrote:
> capacity -> new_capacity?

Done.

[Cutch, 2014-04-21 15:57:19]

Committed patchset #3 manually as r35207 (presubmit successful).