[zeromq-dev] Pull request on atomic ops

Martin Lucina mato at kotelna.sk
Wed May 12 16:53:40 CEST 2010


Guys,

ok, I've reverted the changes in commit f6c1c97.

I also removed the unused native SPARC ops and fixed the asm clobber
statement for atomic_counter::sub() (thanks for catching that).

Steve, your point about the GCC builtins needing implementation is fine but
the problem is that with the current state of affairs there's no easy way
to tell which of the atomic builtins (if any) are implemented on a
particular architecture.

Cheers,

-mato

sustrik at 250bpm.com said:
> Mato, Steven,
> 
> The rule of the thumb should be: Let it work on as many systems as 
> possible -- even with performance penalty.
> 
> Performance penalty is better than not being able to use it at all.
> 
> Once, gcc folks fix the problem, we re-apply the patch.
> 
> Martin
> 
> Steven McCoy wrote:
> > On 11 May 2010 01:04, Martin Lucina <mato at kotelna.sk 
> > <mailto:mato at kotelna.sk>> wrote:
> > 
> >     This is because GCC on ARM does not implement these builtins.  For
> >     details
> >     please refer to my detailed write up on this subject here:
> > 
> > 
> > This only a temporary issue for ARM, post implementation this would be 
> > moot?  Therefore it would be better adding a separate fall back for ARM 
> > rather than promoting it as a primary option.
> > 
> > C++0x includes atomic ops which is why GCC includes the intrinsic API.
> > 
> > http://gcc.gnu.org/wiki/Atomic
> > 
> > -- 
> > Steve-o
> > 
> > 
> > ------------------------------------------------------------------------
> > 
> > _______________________________________________
> > zeromq-dev mailing list
> > zeromq-dev at lists.zeromq.org
> > http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> 
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev



More information about the zeromq-dev mailing list