On Fri 2019-05-03 11:16:26, Tobin C. Harding wrote:
> On Thu, May 02, 2019 at 10:34:12AM +0200, Petr Mladek wrote:
> > On Wed 2019-05-01 09:38:03, Tobin C. Harding wrote:
> > I guess that we need two examples. I currently understand
> > it the following way:
> > 
> > 1. sysfs interface and the structure can be freed anytime:
> > 
> >     struct A
> >     {
> >             struct kobject kobj;
> >             ...
> >     };
> > 
> >     void fn(void)
> >     {
> >             struct A *a;
> >             int ret;
> > 
> >             a = kzalloc(sizeof(*a), GFP_KERNEL);
> >             if (!a)
> >                     return;
> > 
> >             /*
> >              * Initialize structure before we make it accessible via
> >              * sysfs.
> >              */
> >             ret = some_init_fn();
> >             if (ret) {
> >                     goto init_err;
> >             }
> > 
> >             ret = kobject_init_and_add(&a->kobj, ktype, NULL, "foo");
> >             if (ret)
> >                     goto kobj_err;
> > 
> >             return 0;
> > 
> >     kobj_err:
> >             /* kobject_init() always succeds and take reference. */
> >             kobject_put(kobj);
> >             return ret;
> > 
> >     init_err:
> >             /* kobject was not initialized, simple free is enough */
> >             kfree(a);
> >             return ret;
> >     }
> > 
> > 
> > 2. Structure must be registered into the subsystem before
> >    it can be made visible via sysfs:
> > 
> >     struct A
> >     {
> >             struct kobject kobj;
> >             ...
> >     };
> > 
> >     void fn(void)
> >     {
> >             struct A *a;
> >             int ret;
> > 
> >             a = kzalloc(sizeof(*a), GFP_KERNEL);
> >             if (!a)
> >                     return;
> > 
> >             ret = some_init_fn();
> >             if (ret) {
> >                     goto init_err;
> >             }
> > 
> >             /*
> >              * Structure is in a reasonable state and can be freed
> >              * via the kobject release callback.
> >              */
> >             kobject_init(&a->kobj);
> > 
> >             /*
> >              * Register the structure so that it can cooperate
> >              * with the rest of the system.
> >              */
> >             ret = register_fn(a);
> > `           if (ret)
> >                     goto register_err;
> > 
> > 
> >             /* Make it visible via sysfs */
> >             ret = kobject_add(&a->kobj, ktype, NULL, "foo");
> >             if (ret) {
> >                     goto kobj_add_err;
> >             }
> > 
> >             /* Manipulate the structure somehow */
> >             ret = action_fn(a);
> >             if (ret)
> >                     goto action_err;
> > 
> >             mutex_unlock(&my_mutex);
> >             return 0;
> > 
> >     action_err:
> >             /*
> >              * Destroy sysfs interface but the structure
> >              * is still needed.
> >              */
> >             kobject_del(&a->kboject);
> >     kobject_add_err:
> >             /* Make it invisible to the system. */
> >             unregister_fn(a);
> >     register_err:
> >             /* Release the structure unsing the kobject callback */
> >             kobject_put(&a->kobj);
> >             return;
> > 
> >     init_err:
> >             /*
> >              * Custom init failed. Kobject release callback would do
> >              * a double free or so. Simple free is enough.
> >              */
> >              kfree(a);
> >     }
> > 
> > I would really prefer if we clearly understand where each variant makes
> > sense before we start modifying the code and documentation.
> 
> Hi Petr,
> 
> Shall we work these two examples into samples/kobject/.  I'm AFK now for
> the rest of the week but I can do it on Monday if you don't mind me
> doing it?

Sounds good to me. The current samples/kobject/kobject-example shows
the most simple case when the kobject is standalone. While the
above samples shows how to have it bundled in a bigger structure.

Thanks,
Petr

Reply via email to