[ltt-dev] [PATCH 2/2] introduce uatomic_and and uatomic_or

[compudj@krystal.dyndns.org (Mathieu Desnoyers)]

* Paolo Bonzini (pbonzini at redhat.com) wrote:
> On 11/12/2010 01:08 PM, Mathieu Desnoyers wrote:
>> * Paolo Bonzini (pbonzini at redhat.com) wrote:
>>> These are useful to flip single bits.
>>
>> I'm trying to figure out the semantic of these operations.
>>
>> are they
>>
>> uatomic_add / uatomic_or or uatomic_add_return / uatomic_or_return ?
>
> They return the old value, consistently with their names (see also the  
> test_uatomic.c change).  There are three possibilities:
>
> - return nothing
>
> - return the old value
>
> - return the new value
>
> Return nothing can be useful because it can be optimized on x86 as "lock  
> orl (mem), reg/imm".  However, there are no other return-nothing atomic  
> ops in uatomic_*.h so I decided not to provide this.
>
> Returning the new value doesn't make sense for and/or since you cannot  
> revert the operation (unlike uatomic_add which can be implemented from  
> uatomic_add_return).

How about we start by implementing uatomic_and/uatomic_or that return
void, and if we ever need uatomic_return_and/uatomic_return_or (see my
other mail), we add them ?

Otherwise we'll be stucked with heavily synchronized operations to
return a value people just don't care about.

By the way, uatomic_add/uatomic_dec don't return anything.

Mathieu

>
> So I chose the second.
>
> Paolo
>
> _______________________________________________
> ltt-dev mailing list
> ltt-dev at lists.casi.polymtl.ca
> http://lists.casi.polymtl.ca/cgi-bin/mailman/listinfo/ltt-dev
>

-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com