Re: [RFC] [PATCH V2 1/2] input: CMA3000 Accelerometer driver

[Hemanth V]

----- Original Message ----- 
From: "Dmitry Torokhov" <dmitry.torokhov@xxxxxxxxx>
> Hi Hemanth,
>
> On Fri, Sep 03, 2010 at 04:02:11PM +0530, Hemanth V wrote:
> > >
> > >
> > >Do not like repeated release of resources in main and error path... Can
> > >we do it like:
> > >
> > >...
> > >disable_irq();
> > >error = cma3000_set(data, CMA3000_CTRL, ctrl, "ctrl");
> > >if (!error) {
> > >/* Settling time delay required after mode change */
> > >msleep(CMA3000_SETDELAY);
> > >}
> > >enable_irq();
> > >out:
> > >mutex_unlock();
> > >return error ? error : count;
> >
> > I am thinking I can just add the below statement, and should be able to
> > remove repeated release.
> >
> > return error ? error : count;
>
> That would make us sleep for CMA3000_SETDELAY even in case of failure...
> ... But that should be OK.

I suppose it should not sleep if its the modified code segment as below

       error = CMA3000_SET(data, CMA3000_CTRL, ctrl, "ctrl");
       if (error < 0)
               goto err_op3_failed;

       /* Settling time delay required after mode change */
       msleep(CMA3000_SETDELAY);

err_op3_failed:
       enable_irq(data->client->irq);
err_op2_failed:
       mutex_unlock(&data->mutex);
err_op1_failed:
       return (error ? error : count);


> > >>+
> > >>+ if (data->client->irq) {
> > >>+ ret = request_threaded_irq(data->client->irq, NULL,
> > >>+ cma3000_thread_irq,
> > >>+ irqflags | IRQF_ONESHOT,
> > >>+ data->client->name, data);
> > >>+
> > >>+ if (ret < 0) {
> > >>+ dev_err(&data->client->dev,
> > >>+ "request_threaded_irq failed\n");
> > >>+ goto err_op1_failed;
> > >>+ }
> > >>+ }
> > >
> > >What is the utility of the driver when there is no IRQ line?
> >
> > Not sure I fully understand your comments.
> > Currently probe would return a failure.
>
> You have a check for data->client->irq != 0 and finish probe() with
> success in case it is 0. The question is what is the use of the
> device/driver combo in case when data->client->irq == 0?

Yes agreed, will add a error scenario when data->client->irq == 0


--
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/