Re: 2.6.33 GP fault only when built with tracing

[Randy Dunlap <randy.dunlap@oracle.com>]

On 03/23/10 18:20, Mathieu Desnoyers wrote:
> 
> While we wait for the sun to move to other time zones, can you check if the
> following patch fixes your problem ?
> 

Hi Mathieu,
Yes, this works.  Thanks.

I'll test Li's patch now...

> 
> module: fix __module_ref_addr()
> 
> __module_ref_addr() should use per_cpu_ptr() to obfuscate the pointer
> (RELOC_HIDE is needed for per cpu pointers).
> 
> This non-standard per-cpu pointer use has been introduced by commit
> 720eba31f47aeade8ec130ca7f4353223c49170f
> 
> Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
> CC: Eric Dumazet <dada1@cosmosbay.com>
> CC: Rusty Russell <rusty@rustcorp.com.au>
> ---
>  include/linux/module.h |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> Index: linux-2.6-lttng/include/linux/module.h
> ===================================================================
> --- linux-2.6-lttng.orig/include/linux/module.h	2010-03-23 18:11:14.000000000 -0400
> +++ linux-2.6-lttng/include/linux/module.h	2010-03-23 18:14:07.000000000 -0400
> @@ -467,7 +467,7 @@ void symbol_put_addr(void *addr);
>  static inline local_t *__module_ref_addr(struct module *mod, int cpu)
>  {
>  #ifdef CONFIG_SMP
> -	return (local_t *) (mod->refptr + per_cpu_offset(cpu));
> +	return (local_t *) per_cpu_ptr(mod->refptr, cpu);
>  #else
>  	return &mod->ref;
>  #endif


-- 
~Randy