Subj : Re: posix and lock-free algorithms
To   : comp.programming.threads
From : Mayan Moudgill
Date : Fri Aug 12 2005 10:21 pm

David Schwartz wrote:

> 
> "Mayan Moudgill" <mmoudgill@sandbridgetech.com> wrote in message
> news:11fqai13ktc8930@corp.supernews.com...
> 
>  > Actually, it is worse than this. As declared, an ISO standard
>  > compliant C compiler is free to reorder, get rid of, and do
>  > miscellaneous other optimizations to i and j.
> 
>     Of course, but since nobody was talking about an ISO standard 
> compliant  C compiler, how is the an issue?
> 
>  > ISO/ANSI C has *NO* knowledge of multi-threading or concurrent
>  > execution.
> 
>     Right, that's why POSIX was created.
> 
>  > There are only a couple of requirements  that the compiler has to
>  > satisfy:
>  > - the optimized code must have the same black-box behavior as the
>  > abstract
>  > *sequential* machine
>  > - volatiles force order w.r.t. other volatiles.
> 
>     Umm, no, a POSIX-compliant compiler also has to satisfy the 
> requirements imposed by POSIX.

Can you point me to the POSIX standard for the C language?

AFAIK, no such thing exists. Which is why this thread exists - if you 
could define a standard [which would require defining an abstract 
machine] for C, the semantics would be precisely nailed down [well, at 
least the "implementation-defined" and "undefined" stuff would become 
clear]. But since (AFAIK) there is no such beast, I'll say the following 
- since there is no formal semantics for C defined by Posix, Posix 
compliance is a non-sequitur.

.