Re: [PATCH] Free swap suspend from depending upon PageReserved.

[Nigel Cunningham <ncunningham@cyclades.com>]

Hi.

On Wed, 2005-10-05 at 22:12, Pavel Machek wrote:
> Hi!
> 
> > Here's the patch we've previously discussed, which removes the
> > dependancy of swap suspend on PageReserved.
> 
> This ends up in Linus' changelog, so "we've previously discussed"
> is not okay here. Missing signed-off. What is benefit of this?

I wasn't seeking to get it merged immediately, but was seeking comments.

> swsusp part looks okay, but will Andrew like the generic part? I guess
> I'd prefer to postpone this one (unless we are last user of
> PageReserved) -- I do not see too big benefit and there's potential
> for breakage.

There have already been patches to remove work toward removing
PageReserved. If swap suspend isn't the last user, it won't be far from
it.

> > @@ -353,7 +357,7 @@ static void __init pagetable_init (void)
> >  #endif
> >  }
> >  
> > -#ifdef CONFIG_SOFTWARE_SUSPEND
> > +#ifdef CONFIG_PM
> >  /*
> >   * Swap suspend & friends need this for resume because things like the intel-agp
> >   * driver might have split up a kernel 4MB mapping.
> 
> This is wrong, right? It 

No. I use it too. From your perspective though, I suppose it is wrong.

> > @@ -540,6 +544,7 @@ void __init mem_init(void)
> >  	int codesize, reservedpages, datasize, initsize;
> >  	int tmp;
> >  	int bad_ppro;
> > +	void * addr;
> 
> Please make it void *addr;.

Ok.

If noone else suggests changes, I'll properly submit it.

Regards,

Nigel

> 								Pavel
--