Re: [this_cpu_xx V8 07/16] Module handling: Use this_cpu_xx to dynamically allocate counters

[Tejun Heo <tj@kernel.org>]

Hello,

On 12/19/2009 07:26 AM, Christoph Lameter wrote:
> Use cpu ops to deal with the per cpu data instead of a local_t. Reduces memory
> requirements, cache footprint and decreases cycle counts.
> 
> The this_cpu_xx operations are also used for !SMP mode. Otherwise we could
> not drop the use of __module_ref_addr() which would make per cpu data handling
> complicated. this_cpu_xx operations have their own fallback for !SMP.
> 
> The last hold out of users of local_t is the tracing ringbuffer after this patch
> has been applied.
> 
> Signed-off-by: Christoph Lameter <cl@linux-foundation.org>

Please keep Rusty Russell cc'd on module changes.

> Index: linux-2.6/kernel/trace/ring_buffer.c
> ===================================================================
> --- linux-2.6.orig/kernel/trace/ring_buffer.c	2009-12-18 13:13:24.000000000 -0600
> +++ linux-2.6/kernel/trace/ring_buffer.c	2009-12-18 14:15:57.000000000 -0600
> @@ -12,6 +12,7 @@
>  #include <linux/hardirq.h>
>  #include <linux/kmemcheck.h>
>  #include <linux/module.h>
> +#include <asm/local.h>

This doesn't belong to this patch, right?  I stripped this part out,
added Cc: to Rusty and applied 1, 2, 7 and 8 to percpu branch.  I'll
post the updated patch here.

Thanks.

-- 
tejun