Re: [boost] [interprocess] atomic_write32 on GCC/x86/x68_64 lacks memory barrier

Subject: Re: [boost] [interprocess] atomic_write32 on GCC/x86/x68_64 lacks memory barrier
From: Peter Dimov (lists_at_[hidden])
Date: 2014-08-11 10:19:11

Lars Hagström wrote:
> Hmm, should I add a fence to the Windows atomic_read32 variant too?
> Reading the Boost.Atomics code it looks like it puts a _ReadWriteBarrier 
> in there...

That would be a good idea as well. Some MSVC versions give volatile reads 
acquire semantics, and some do not, so it's safer to have the barrier.