Re: [PATCH 6/7] module: add two missing synchronize_sched()

[Eric Dumazet]

On Thu, 2012-03-15 at 22:49 +0800, Cong Wang wrote:
> In RCU doc, it said synchronize_sched() is used to mark the end
> of update, so two synchronize_sched() are missing after
> module list add/del.
> 
> Cc: Eric Dumazet <eric.dumazet@xxxxxxxxx>
> Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx>
> Cc: Rusty Russell <rusty@xxxxxxxxxxxxxxx>
> Signed-off-by: Cong Wang <xiyou.wangcong@xxxxxxxxx>
> ---
>  kernel/module.c |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/kernel/module.c b/kernel/module.c
> index 06e7dc5..2e7a5c5 100644
> --- a/kernel/module.c
> +++ b/kernel/module.c
> @@ -1775,6 +1775,7 @@ static void free_module(struct module *mod)
>  	mutex_lock(&module_mutex);
>  	stop_machine(__unlink_module, mod, NULL);
>  	mutex_unlock(&module_mutex);
> +	synchronize_sched();
>  	mod_sysfs_teardown(mod);
>  
>  	/* Remove dynamic debug info */
> @@ -2928,6 +2929,7 @@ static struct module *load_module(void __user *umod,
>  	module_bug_finalize(info.hdr, info.sechdrs, mod);
>  	list_add_rcu(&mod->list, &modules);
>  	mutex_unlock(&module_mutex);
> +	synchronize_sched();
>  
>  	/* Module is ready to execute: parsing args may do that. */
>  	err = parse_args(mod->name, mod->args, mod->kp, mod->num_kp, NULL);

Have a look at stop_machine() first

after list_add_rcu(), there is absolutely no such requirement.



--
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/