mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joseph Wu" <jos...@mesosphere.io>
Subject Re: Review Request 39949: Document and simplify libprocess initialization synchronization logic.
Date Wed, 04 Nov 2015 19:52:58 GMT


> On Nov. 4, 2015, 11:25 a.m., Neil Conway wrote:
> > Can we test that more thoroughly than just "make check"? e.g., if there's a unit
test that tries to enter this logic with multiple threads at once, running that with gtest_repeat=1000
would be nice.

There are tons of methods in libprocess that call `process::initialize` as a side-effect,
but at the same time, the libprocess test suite starts out with an essentially race-free init
call (See: https://github.com/apache/mesos/blob/master/3rdparty/libprocess/src/tests/main.cpp#L52).
 So any `--gtest_repeat` or `--gtest_shuffle` won't actually test the init code.  (The master
and agent also call init once on startup.)

I'm not sure how valuable it will be to, say, to spawn a bunch of threads that call `process::initialize`.
 Do you have any suggestions?


> On Nov. 4, 2015, 11:25 a.m., Neil Conway wrote:
> > 3rdparty/libprocess/src/process.cpp, line 742
> > <https://reviews.apache.org/r/39949/diff/1/?file=1115784#file1115784line742>
> >
> >     I wonder if it would be an improvement to rename these variables to reflect
what they are used for more clearly.
> >     
> >     For example: maybe call them "initialize_started" and "initialize_complete"
(and change the second so that it goes from false -> true instead of true -> false).

That sounds very reasonable.  I'll rename (and see if anyone objects :).


- Joseph


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/39949/#review105115
-----------------------------------------------------------


On Nov. 4, 2015, 10:58 a.m., Joseph Wu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/39949/
> -----------------------------------------------------------
> 
> (Updated Nov. 4, 2015, 10:58 a.m.)
> 
> 
> Review request for mesos, Benjamin Hindman and Joris Van Remoortere.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> The initialization synchronization logic contains three conditions, which check:
> 1) Was `initialize` called and is it done?
> 2) Was `initialize` called and is it not done?
> 3) Are you the first to call `initialize`?
> 
> Condition (3) uses `compare_exchange_strong` between `initialized` and `false`.  This
returns `true` (and sets `initialized` to true) iff the caller is the first to reach that
expression.
> 
> The second simultaneous caller of `initialize` will either satisify condition (2) or
(3) and then wait on `initializing`.  For the second caller, (2) and (3) are identical because
`compare_exchange_strong` between `true` and `false` will always return false, thereby putting
the second caller into the waiting loop.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/src/process.cpp a94712b9ac3b60fb047b3a5a4d84a56fa4d02313 
> 
> Diff: https://reviews.apache.org/r/39949/diff/
> 
> 
> Testing
> -------
> 
> `make check`
> 
> 
> Thanks,
> 
> Joseph Wu
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message