[M3commit] CVS Update: cm3

[Tony Hosking]

It's fine if the result is stored in memory.  We can pick it up from there, right?  Rereading memory distinct from the atomic variable is fine.

On 12 Feb 2010, at 14:36, Jay Krell wrote:

> CVSROOT:	/usr/cvs
> Changes by:	jkrell at birch.	10/02/12 14:36:19
> 
> Modified files:
> 	cm3/m3-sys/m3back/src/: M3x86.m3 
> 
> Log message:
> 	nevermind, this is all wrong except for store, load, fence, exchange,
> 	and compare_exchange
> 	The atomic interface really does *not* expose a bunch of
> 	functionality on x86. x86 generally stores the result in memory.
> 	We cannot then return the result because we'd have to reread memory.
> 	Therefore all these functions will require cmpxchg loops.
> 	Probably just write wrapper functions for the C compiler intrinsics..