Simplify timeline backend

Simplify timeline backend

Now that we have a central list of all threads that are interacting with the VM we can stop relying on the Isolate's thread registry to cache timeline blocks.

- Move from caching blocks in the Isolate's thread registry and caching them in the Thread directly.
- This allows cached blocks to have events from multiple isolates, increasing usage density and reducing the number of blocks being cached at any given time.
- We no longer need a distinct global block, this allows me to remove a bunch of dead code.

R=turnidge@google.com

Committed: https://github.com/dart-lang/sdk/commit/da946fbf99dc9a12d37bf0166f88c00f9d4f5564

[Cutch, 2015-10-16 18:21:19]

johnmccutchan@google.com changed reviewers:
+ turnidge@google.com

[turnidge, 2015-10-16 19:40:20]

lgtm.  2 small questions.

Gosh, this is an improvement.

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc
File runtime/vm/timeline.cc (right):

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc#newc...
runtime/vm/timeline.cc:44: // a |TimelineEventBlock| cached in TLS partially
filled with events. In order
Is it really cached in TLS?  Seems out of date.

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc#newc...
runtime/vm/timeline.cc:135: thread->set_timeline_block(NULL);
Should we release the timeline_block_lock before we call FinishBlock?  Or is it
irrelevant?

[Cutch, 2015-10-16 19:43:52]

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc
File runtime/vm/timeline.cc (right):

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc#newc...
runtime/vm/timeline.cc:44: // a |TimelineEventBlock| cached in TLS partially
filled with events. In order
On 2015/10/16 19:40:20, turnidge wrote:
> Is it really cached in TLS?  Seems out of date.

Done.

https://codereview.chromium.org/1402383003/diff/1/runtime/vm/timeline.cc#newc...
runtime/vm/timeline.cc:135: thread->set_timeline_block(NULL);
On 2015/10/16 19:40:20, turnidge wrote:
> Should we release the timeline_block_lock before we call FinishBlock?  Or is
it
> irrelevant?

For now I'm keeping this consistent with the documented lock ordering. In the
future I would like to do this. I've added a TODO for myself to circle back.

[Cutch, 2015-10-16 20:37:03]

Committed patchset #3 (id:40001) manually as
da946fbf99dc9a12d37bf0166f88c00f9d4f5564 (presubmit successful).