Re: [boost] This deadlocks on Linux but not on Windows

From: John Torjo (john.groups_at_[hidden])
Date: 2008-02-07 07:44:19

Jos Hickson wrote:
> On 07/02/2008, Clark Sims <clark_sims_boost_at_[hidden]> wrote:
>   
>> I would like it to deadlock on windows, because I just wrote a large header file, which assumes that a deadlock will occur.  In fact, I don't know how I am going to program mutex's and locks,  if they act like recursive locks on windows, but scoped locks on linux.
>>
>>     
>
> We've noticed this variation behaviour in the past when using
> boost::mutex.  Under the hood on Windows it uses a critical section
> which I believe is recursive.  Not sure if the rewritten version of
> Boost.Threads will do the same.
>
>   
Yes, critical sections are recursive.
I looked a bit at this: http://rafb.net/p/u0LDDP45.html
Did not look into too much detail, *but*
you can definitely make it work even in the presence of recursive mutexes.

Best,
John
> Sorry, but all I can do is confirm the behaviour.  I believe the
> Boost.Threads concept documentation for mutex and scoped_lock say that
> attempting a recursive lock gives "undefined behaviour" rather than
> definitely deadlocking.
>
> Regards,
>
> Jos
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>
>   

-- 
http://John.Torjo.com -- C++ expert
http://blog.torjo.com
... call me only if you want things done right