Re: 2.6.14-rc1-git-now still dying in mm/slab - this time line 1849

[Andrew Morton <akpm@osdl.org>]

Ravikiran G Thirumalai <kiran@scalex86.org> wrote:
>
> On Thu, Sep 29, 2005 at 06:43:05PM +0200, Petr Vandrovec wrote:
> > Ravikiran G Thirumalai wrote:
> > 
> > Unfortunately I must confirm that it does not fix problem.  But it pointed
> > out to me another thing - proc_inode_cache stuff is put into caches
> > BEFORE this code is executed.  So if anything in mm/slab.c relies
> > on node_to_mask[] being valid (and if it relies on some other things
> > which are set this late), it probably won't work.
> 
> The tests Alok carried out on Petr's box confirmed that cpu_to_node[BP] 
> is not setup early enough by numa_init_array due to the x86_64 changes in
>  2.6.14-rc*, and unfortunately set wrongly by the work around code in 
> numa_init_array(). cpu_to_node[0] gets set with 1 early and later gets set 
> properly to 0 during identify_cpu() when all cpus are brought up, but 
> confusing the numa slab in the process.
> 
> Here is a quick fix for this.  The right fix obviously is to have
> cpu_to_node[bsp] setup early for numa_init_array().  The following patch
> will fix the problem now, and the code can stay on even when cpu_to_node{BP] 
> gets fixed early correctly.
> 
> Thanks to Petr for access to his box.
> 
> Signed off by: Ravikiran Thirumalai <kiran@scalex86.org>
> Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
> 
> Index: slab-x86_64-fix-2.6.14-rc2/arch/x86_64/mm/numa.c
> ===================================================================
> --- slab-x86_64-fix-2.6.14-rc2.orig/arch/x86_64/mm/numa.c	2005-09-29 20:39:25.000000000 -0700
> +++ slab-x86_64-fix-2.6.14-rc2/arch/x86_64/mm/numa.c	2005-09-29 21:38:05.000000000 -0700
> @@ -167,15 +167,14 @@
>  	   mapping. To avoid this fill in the mapping for all possible
>  	   CPUs, as the number of CPUs is not known yet. 
>  	   We round robin the existing nodes. */
> -	rr = 0;
> +	rr = first_node(node_online_map);
>  	for (i = 0; i < NR_CPUS; i++) {
>  		if (cpu_to_node[i] != NUMA_NO_NODE)
>  			continue;
> +		cpu_to_node[i] = rr;
>  		rr = next_node(rr, node_online_map);
>  		if (rr == MAX_NUMNODES)
>  			rr = first_node(node_online_map);
> -		cpu_to_node[i] = rr;
> -		rr++; 
>  	}
>  
>  }

Is this fix independent from
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.14-rc2/2.6.14-rc2-mm2/broken-out/x86_64-fix-the-bp-node_to_cpumask.patch
?


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