Re: [RFC] Introduce atomic_long_t

[Andi Kleen <ak@suse.de>]

On Fri, Dec 09, 2005 at 11:33:28PM +0100, Adrian Bunk wrote:
> On Fri, Dec 09, 2005 at 11:20:45PM +0100, Andi Kleen wrote:
> > > I'd say the sequence is:
> > > 1. create an linux/atomic.h the #include's asm/atomic.h
> > > 2. convert all asm/atomic.h to use linux/atomic.h
> > > 3. move common code to linux/atomic.h
> > 
> > I don't think there is much common code actually. atomic_t 
> > details vary widly between architectures. Just defining
> > a few macros to others is really not significant. I think 
> > Christoph's original patch was just fine.
> 
> All of Christoph's original patch contains common code.
> 
> The amount of duplication his patch would create alone would IMHO be 
> worth creating an linux/atomic.h.

There wasn't actually much code in there. And defining 
asm-generic/atomic-long-on-32bit.h and asm-generic/atomic-long-on-64bit.h
like you essentially proposed would just obfuscate the code, not make it 
easier to maintain.

Aiming for common code is ok, but only when it actually improves
maintainability. 

-Andi

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>