From: Jacob Keller <jacob.keller@gmail.com>
Date: Fri, 28 Oct 2016 18:49:25 GMT
Subject: Re: [PATCH] compat: Allow static initializer for pthreads on Windows
Message-ID: <CA+P7+xoNSrhaQ50GZ4RzLregznw+0yf2=ttNuYf1VUw+YabdFQ@mail.gmail.com>
In-Reply-To: <B5FD05E58E36480894F1BDBBC9589EE1@PhilipOakley>

On Fri, Oct 28, 2016 at 6:01 AM, Philip Oakley <philipoakley@iee.org> wrote:
> From: "Johannes Sixt" <j6t@kdbg.org>
>>
>>
>> One point is that the DCLP idiom must be implemented correctly. There are
>> solutions, of course, and when the initialization is over, we have a
>> miniscule overhead at each pthread_mutex_lock call.
>>
>
> I had to look up DCLP ( = Double Checked Locking Patterns), and found a good
> write up on the issues..
>
> http://www.aristeia.com/Papers/DDJ_Jul_Aug_2004_revised.pdf "C++ and the
> Perils of Double-Checked Locking", which include 'C' issues, and
> multi-thread, multi-processor issues. Not an easy issue when fighting
> optimisers..
>
> --
>
> Philip


Yep, this is why we have memory barriers. Ofcourse languages like C
don't really allow you to express them in the language and we restore
to various platform specific methods.

Thanks,
Jake