Re: [patch] x86, mm: Clean up initmem_init

[Ingo Molnar <mingo@elte.hu>]

* David Rientjes <rientjes@google.com> wrote:

> On Thu, 3 Mar 2011, Yinghai Lu wrote:
> 
> > he want
> > 
> > > +#ifdef CONFIG_ACPI_NUMA
> > >> > +		ret = numa_init(x86_acpi_numa_init);
> > >> > +		if (!ret)
> > >> > +			return;
> > >> > +#endif
> > >> > +#ifdef CONFIG_AMD_NUMA
> > >> > +		ret = numa_init(amd_numa_init);
> > >> > +		if (!ret)
> > >> > +			return;
> > >> > +#endif
> > 
> > to be replaced by:
> > 
> > > +#ifdef CONFIG_ACPI_NUMA
> > >> > +		if (!numa_init(x86_acpi_numa_init))
> > >> > +			return;
> > >> > +#endif
> > >> > +#ifdef CONFIG_AMD_NUMA
> > >> > +		if (!numa_init(amd_numa_init))
> > >> > +			return;
> > >> > +#endif
> 
> It's a matter of style and I think it's up to Ingo what he'd prefer to 
> see.

I think your variant is cleaner: hiding function call side-effects in conditionscan 
be a fragile thing to do. We want constant expressions with no side-effects - so if 
functions are called they should be constant functions as well.

Code compactness isn't everything - if it was we'd be using C to the max to create 
unreadable compound expressions all the time.

Thanks,

	Ingo