Re: [problem] raid performance loss with 2.6.26-rc8 on 32-bit x86(bisected)

[Linus Torvalds]

On Thu, 3 Jul 2008, Mel Gorman wrote:

> diff -rup -X /usr/src/patchset-0.6/bin//dontdiff linux-2.6.26-rc8-clean/mm/page_alloc.c linux-2.6.26-rc8-fix-kswapd-on-numa/mm/page_alloc.c
> --- linux-2.6.26-rc8-clean/mm/page_alloc.c	2008-06-24 18:58:20.000000000 -0700
> +++ linux-2.6.26-rc8-fix-kswapd-on-numa/mm/page_alloc.c	2008-07-02 21:14:16.000000000 -0700
> @@ -2328,7 +2328,8 @@ static void build_zonelists(pg_data_t *p
>  static void build_zonelist_cache(pg_data_t *pgdat)
>  {
>  	pgdat->node_zonelists[0].zlcache_ptr = NULL;
> -	pgdat->node_zonelists[1].zlcache_ptr = NULL;
> +	if (NUMA_BUILD)
> +		pgdat->node_zonelists[1].zlcache_ptr = NULL;
>  }

This makes no sense.

That whole thing is inside a

	#ifdef CONFIG_NUMA
	... numa code ..
	#else
	... this code ..
	#endif

so CONFIG_NUMA will _not_ be set, and NUMA_BUILD is always 0.

So why do that

	if (NUMA_BUILD)
		..

at all, when it is known to be false?

So the patch may be correct, but wouldn't it be better to just remove the 
line entirely, instead of moving it into a conditional that cannot be 
true?

Also, I'm not quite seeing why those zonelists should be zeroed out at 
all. Shouldn't a non-NUMA setup always aim to have node_zonelists[0] == 
node_zonelists[1] == all appropriate zones?

I have to say, the whole mmzoen thing is confusing. The code makes my eyes 
bleed. I can't really follow it.

		Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/