Re: [PATCH] module: add kset_obj_exists() and use it

[Rusty Russell]

Veaceslav Falico <vfal...@redhat.com> writes:
> On Wed, Apr 10, 2013 at 04:47:34PM +0930, Rusty Russell wrote:
>>That's a bug.  We should be cleaning up sysfs before we unlike the
>>removed module from the list.
>>
>>Because the same thing applies to ddebug info, which is also keyed by
>>module name.
>>
>>Something like this (untested!):
>
> Sorry for the late response - I wanted to test it for a longer time.
>
> Your patch works flawlessly and fixes this race, with just a small
> addition, cause otherwise we could BUG() in show_initstate().
>
> Can you apply this patch or should I (re-)send it somehow?
>
> Thank you!
>
> diff --git a/kernel/module.c b/kernel/module.c
> index d0afe23..8be6e97 100644
> --- a/kernel/module.c
> +++ b/kernel/module.c
> @@ -1063,6 +1063,9 @@ static ssize_t show_initstate(struct module_attribute 
> *mattr,
>       case MODULE_STATE_GOING:
>               state = "going";
>               break;
> +     case MODULE_STATE_UNFORMED:
> +             state = "unformed";
> +             break;
>       default:
>               BUG();
>       }

Prefer to remove from sysfs before marking it unformed, like so:

diff --git a/kernel/module.c b/kernel/module.c
index 2468fda..2e7189f 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1862,12 +1862,12 @@ static void free_module(struct module *mod)
 {
        trace_module_free(mod);
 
-       /* We leave it in list to prevent duplicate loads while we clean
-        * up sysfs, ddebug and any other external exposure. */
-       mod->state = MODULE_STATE_UNFORMED;
-
        mod_sysfs_teardown(mod);
 
+       /* We leave it in list to prevent duplicate loads, but make sure
+        * that noone uses it while it's being deconstructed. */
+       mod->state = MODULE_STATE_UNFORMED;
+
        /* Remove dynamic debug info */
        ddebug_remove_module(mod->name);
 
Here's the updated total patch:

diff --git a/kernel/module.c b/kernel/module.c
index 69d2600..2e7189f 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1862,12 +1862,12 @@ static void free_module(struct module *mod)
 {
        trace_module_free(mod);
 
-       /* Delete from various lists */
-       mutex_lock(&module_mutex);
-       stop_machine(__unlink_module, mod, NULL);
-       mutex_unlock(&module_mutex);
        mod_sysfs_teardown(mod);
 
+       /* We leave it in list to prevent duplicate loads, but make sure
+        * that noone uses it while it's being deconstructed. */
+       mod->state = MODULE_STATE_UNFORMED;
+
        /* Remove dynamic debug info */
        ddebug_remove_module(mod->name);
 
@@ -1880,6 +1880,11 @@ static void free_module(struct module *mod)
        /* Free any allocated parameters. */
        destroy_params(mod->kp, mod->num_kp);
 
+       /* Now we can delete it from the lists */
+       mutex_lock(&module_mutex);
+       stop_machine(__unlink_module, mod, NULL);
+       mutex_unlock(&module_mutex);
+
        /* This may be NULL, but that's OK */
        unset_module_init_ro_nx(mod);
        module_free(mod, mod->module_init);

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