Re: [PATCH update5] drivers: add LCD support

["Miguel Ojeda" <maxextreme@gmail.com>]

On 10/31/06, Andrew Morton <akpm@osdl.org> wrote:
> On Wed, 1 Nov 2006 00:37:48 +0000
> Miguel Ojeda Sandonis <maxextreme@gmail.com> wrote:
>
> > Andrew, here it is the fifth update for the drivers-add-lcd-support.
>
> It does queue_delayed_work(), but forgot to do cancel_delayed_work().  If
> the timer is still pending after module unload, deadly things will happen.
>

Hum, I call cancel_delayed_work() as well as flush_workqueue() at module exit:

static void __exit cfag12864b_exit(void)
{
...
+       cfag12864b_disable();
...
}

And cfag12864b_disable() will call them, as fast as the mutex is
unlocked (in fact, modules must call cfag12864b_disable() when they
finish working with the LCD, as cfag12864bfb, for example, does, but
just for be safe, I called it at module unload in cfag12864b. If it is
already disabled, it won't hurt in any way.)

Is that right?

> There's a lot of "if(" in there.  Usual kernel style is "if (".
>

Excuse me, will fix in a few moments.