Re: starvation in pthread_once?

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: starvation in pthread_once?

Gottlob Frege
Blast from the past - whatever happened to changing call_once() to not
create the named mutex when it wasn't needed?


On Tue, Mar 22, 2005 at 1:26 PM, Gottlob Frege <[hidden email]> wrote:

> If it comes down to it, I might vote for daisy-chaining over
> busy-looping (assuming the busy-looping is endless).  Remember, this
> all started because the original implementation was polling/sleeping
> on 'initted' - and if the busy-looping thread is high-priority, then
> we are locked forever...
>
>
> On Tue, 22 Mar 2005 15:14:07 +1100, Ross Johnson
> <[hidden email]> wrote:
>> On Mon, 2005-03-21 at 11:07 -0500, Gottlob Frege wrote:
>>
>> > So, it doesn't seem to be getting any easier!  *Almost* to the point
>> > where a big named mutex becomes tempting - there is a lot to be said
>> > for simplicity.  However, my/the goal is still to at least minimize
>> > the non-contention simple init case...
>>
>> I'm less and less tempted to use a named mutex. Perhaps there's a
>> standard technique, but AFAICS it's impossible to guarantee that the
>> name is unique across the system (and all Windows variants).
>>
>> And I agree, minimum overhead for the uncontended case is the top
>> priority (after correct behaviour). I'm not concerned at all about speed
>> in the cancellation case.
>>
>> > And the event is still an auto-reset, although I no longer think it
>> > really matters - I really haven't had the tenacity to think this stuff
>> > through.  If it doesn't matter, manual-reset would be better, I think
>> > - I don't like having 1 thread relying on another thread waking it up,
>> > - for cases where the thread is killed, or strange thread priorities,
>> > etc.
>>
>> It all looks to me like it will work. I don't recall, in the version
>> that's in pthreads-win32 now, why I included eventUsers (++/--) in what
>> you have as the __lock() sections. Maybe to save additional Atomic calls
>> (bus locks). But now I realise [in that version - not yours] that waking
>> threads can block unnecessarily when leaving the wait section.
>>
>> It probably doesn't matter if cancel_event is auto or manual. I think
>> there will be at most one thread waiting on it. And, for 'event', like
>> you I'm uncomfortable with daisy-chaining SetEvent() calls.
>>
>> The only problem with the alternative of using a manual-reset event is
>> that some thread/s may busy-loop for a bit until an explicit reset
>> occurs. It seems untidy, but it's probably more robust than daisy-
>> chained SetEvents given the issues you've identified above.
>>
>> So I'm tempted to leave both events as manual-reset events. I'm also
>> guessing that this busy-looping will be extremely rare - perhaps only
>> when a new thread sneaks in to become initter, then suspends just inside
>> while the first waiter is waking and heading back to the loop start.
>>
>> I'll run your design and let you know the results.
>>
>> Thanks.
>> Ross
>>
>>
>
Reply | Threaded
Open this post in threaded view
|

Re: starvation in pthread_once?

Ross Johnson-2
Hi Gottlob,

Vladimir Kliatchko reimplemented pthread_once to use his
implementation of MCS (Mellor-Crummey/Scott) locks.

 From ptw32_MCS_lock.c:-

 * About MCS locks:
 *
 * MCS locks are queue-based locks, where the queue nodes are local to the
 * thread. The 'lock' is nothing more than a global pointer that points to
 * the last node in the queue, or is NULL if the queue is empty.
 *
 * Originally designed for use as spin locks requiring no kernel resources
 * for synchronisation or blocking, the implementation below has adapted
 * the MCS spin lock for use as a general mutex that will suspend threads
 * when there is lock contention.
 *
 * Because the queue nodes are thread-local, most of the memory read/write
 * operations required to add or remove nodes from the queue do not trigger
 * cache-coherence updates.
 *
 * Like 'named' mutexes, MCS locks consume system resources transiently -
 * they are able to acquire and free resources automatically - but MCS
 * locks do not require any unique 'name' to identify the lock to all
 * threads using it.



Gottlob Frege wrote:

> Blast from the past - whatever happened to changing call_once() to not
> create the named mutex when it wasn't needed?
>
>
> On Tue, Mar 22, 2005 at 1:26 PM, Gottlob Frege <[hidden email]> wrote:
>  
>> If it comes down to it, I might vote for daisy-chaining over
>> busy-looping (assuming the busy-looping is endless).  Remember, this
>> all started because the original implementation was polling/sleeping
>> on 'initted' - and if the busy-looping thread is high-priority, then
>> we are locked forever...
>>
>>
>> On Tue, 22 Mar 2005 15:14:07 +1100, Ross Johnson
>> <[hidden email]> wrote:
>>    
>>> On Mon, 2005-03-21 at 11:07 -0500, Gottlob Frege wrote:
>>>
>>>      
>>>> So, it doesn't seem to be getting any easier!  *Almost* to the point
>>>> where a big named mutex becomes tempting - there is a lot to be said
>>>> for simplicity.  However, my/the goal is still to at least minimize
>>>> the non-contention simple init case...
>>>>        
>>> I'm less and less tempted to use a named mutex. Perhaps there's a
>>> standard technique, but AFAICS it's impossible to guarantee that the
>>> name is unique across the system (and all Windows variants).
>>>
>>> And I agree, minimum overhead for the uncontended case is the top
>>> priority (after correct behaviour). I'm not concerned at all about speed
>>> in the cancellation case.
>>>
>>>      
>>>> And the event is still an auto-reset, although I no longer think it
>>>> really matters - I really haven't had the tenacity to think this stuff
>>>> through.  If it doesn't matter, manual-reset would be better, I think
>>>> - I don't like having 1 thread relying on another thread waking it up,
>>>> - for cases where the thread is killed, or strange thread priorities,
>>>> etc.
>>>>        
>>> It all looks to me like it will work. I don't recall, in the version
>>> that's in pthreads-win32 now, why I included eventUsers (++/--) in what
>>> you have as the __lock() sections. Maybe to save additional Atomic calls
>>> (bus locks). But now I realise [in that version - not yours] that waking
>>> threads can block unnecessarily when leaving the wait section.
>>>
>>> It probably doesn't matter if cancel_event is auto or manual. I think
>>> there will be at most one thread waiting on it. And, for 'event', like
>>> you I'm uncomfortable with daisy-chaining SetEvent() calls.
>>>
>>> The only problem with the alternative of using a manual-reset event is
>>> that some thread/s may busy-loop for a bit until an explicit reset
>>> occurs. It seems untidy, but it's probably more robust than daisy-
>>> chained SetEvents given the issues you've identified above.
>>>
>>> So I'm tempted to leave both events as manual-reset events. I'm also
>>> guessing that this busy-looping will be extremely rare - perhaps only
>>> when a new thread sneaks in to become initter, then suspends just inside
>>> while the first waiter is waking and heading back to the loop start.
>>>
>>> I'll run your design and let you know the results.
>>>
>>> Thanks.
>>> Ross
>>>
>>>
>>>