Re: [BUG] at mm/slab.c:3320

[KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

On Mon, 7 Jan 2008 10:10:16 -0800 (PST)
Christoph Lameter <clameter@sgi.com> wrote:

> On Mon, 7 Jan 2008, KAMEZAWA Hiroyuki wrote:
> 
> > Seems Node 1 has no NORMAL memory.
> > 
> > Because the patch changes 'online_node' to N_NORMAL_MEMORY, there is a change.
> > I'm not sure but cachep->nodelists[] should be created against all online nodes ?
> 
> Well what is the point of creating a memory structure for a node from 
> which no memory for the slab allocator can be allocated? I think we need a 
> special cpu_to_node() that only takes normal memory into consideration. 
> 
In usual alloc_pages() allocator, this is done by zonelist fallback.

> And we need to use that new function (cpu_to_node_normal_memory or so?) to 
> find memory for the slab and other stuff in the kernel.
> 

It seems that cache->nodelists[nid] == NULL case should be handled even if
nid == cpu_to_node(smp_processor_id()). 

complicated ?

Thanks,
-Kame

--
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>