volatile qualifier hurts single-threaded optimized case

[Paolo Carlini]

Benjamin Kosnik wrote:

>>>  static inline _Atomic_word
>>>  __attribute__ ((__unused__))
>>>-  __exchange_and_add_dispatch(volatile _Atomic_word* __mem, int __val)
>>>+  __exchange_and_add_dispatch(_Atomic_word* __mem, int __val)
>>> 
>>>
>>>      
>>>
>>Agreed, but doing this means that in the improvements to rope that 
>>Richard (and Hans) also want we have to call __exchange_and_add and 
>>__atomic_add directly,  because in that case the involved memory 
>>locations are volatile and have to remain like that per Hans desiderata, 
>>I gather. 
>>    
>>
>I find this somewhat confusing, but am going to go ahead with this, as
>I have other fixes in the pipeline for this file.
>  
>
Agreed, agreed. I was just pointing out, for completeness, that we had a 
very long discussion thread, Hans also participating: in that thread 
Richard proposed to change rope to use the atomic functions and prepared 
a patch. Hans was very happy with that patch besides the delicate point 
that, because of races in some cases (i.e., accesses to the atomic 
memory *not* using the primitives for performance reasons) the involved 
memory had to remain volatile for now, to be more safe. Then, the 
dispatchers could not be used...

Paolo.