Update WaitSetDispatcher to use the new persistent Awakable stuff.

Update WaitSetDispatcher to use the new persistent Awakable stuff.

* Now the wait set accurately tracks the state of everything it watches.
* And there's no more need for the "busy" stuff.
* I still have to implement MojoWaitSetWait(), but that should actually
  be much more straightforward now that state is accurately tracked.

R=vardhan@google.com
BUG=#350

Committed: https://chromium.googlesource.com/external/mojo/+/17e05c2be2a0c46f2ec88d49626af6770361b5cd

[vardhan, 2016-06-23 18:26:11]

lgtm w/ question

https://codereview.chromium.org/2085373003/diff/1/mojo/edk/system/wait_set_di...
File mojo/edk/system/wait_set_dispatcher.cc (right):

https://codereview.chromium.org/2085373003/diff/1/mojo/edk/system/wait_set_di...
mojo/edk/system/wait_set_dispatcher.cc:299:
!entry->signals_state.can_satisfy(entry->signals)) {
is this right? if entry's signal state cannot satisfy the desired signal, we
still try to trigger?

(sorry, i might be missing something)

[viettrungluu, 2016-06-23 20:01:06]

Thanks.

https://codereview.chromium.org/2085373003/diff/1/mojo/edk/system/wait_set_di...
File mojo/edk/system/wait_set_dispatcher.cc (right):

https://codereview.chromium.org/2085373003/diff/1/mojo/edk/system/wait_set_di...
mojo/edk/system/wait_set_dispatcher.cc:299:
!entry->signals_state.can_satisfy(entry->signals)) {
On 2016/06/23 18:26:11, vardhan wrote:
> is this right? if entry's signal state cannot satisfy the desired signal, we
> still try to trigger?

Yes, we want to report that it's become "never"-satisfiable (via a
MOJO_RESULT_FAILED_PRECONDITION in the wait_result field of the
MojoWaitSetResult).

This is analogous to the existing Wait/WaitMany.

> 
> (sorry, i might be missing something)

[viettrungluu, 2016-06-23 20:02:07]

Description was changed from

==========
Update WaitSetDispatcher to use the new persistent Awakable stuff.

* Now the wait set accurately tracks the state of everything it watches.
* And there's no more need for the "busy" stuff.
* I still have to implement MojoWaitSetWait(), but that should actually
  be much more straightforward now that state is accurately tracked.

R=vardhan@google.com
BUG=#350
==========

to

==========
Update WaitSetDispatcher to use the new persistent Awakable stuff.

* Now the wait set accurately tracks the state of everything it watches.
* And there's no more need for the "busy" stuff.
* I still have to implement MojoWaitSetWait(), but that should actually
  be much more straightforward now that state is accurately tracked.

R=vardhan@google.com
BUG=#350

Committed:
https://chromium.googlesource.com/external/mojo/+/17e05c2be2a0c46f2ec88d49626...
==========

[viettrungluu, 2016-06-23 20:02:10]

Committed patchset #1 (id:1) manually as
17e05c2be2a0c46f2ec88d49626af6770361b5cd (presubmit successful).