Re: [PATCH v2 01/24] input: Add ff-memless-next module

["Michal Malý" <madcatxster@devoid-pointer.net>]

On Wednesday 23 of April 2014 14:12:59 Oliver Neukum wrote:
> On Tue, 2014-04-22 at 15:59 +0200, Michal Malý wrote:
> > +/* Some devices might have a limit on how many uncombinable effects
> > + * can be played at once */
> > +static int mlnx_upload_conditional(struct mlnx_device *mlnxdev,
> > +                                  const struct ff_effect *effect)
> > +{
> > +       struct mlnx_effect_command ecmd = {
> > +               .cmd = MLNX_UPLOAD_UNCOMB,
> > +               .u.uncomb.id = effect->id,
> > +               .u.uncomb.effect = effect
> > +       };
> > +       return mlnxdev->control_effect(mlnxdev->dev, mlnxdev->private,
> > &ecmd);
> > +}
> > +
> 
> This mean you are building the structure on the stack
> 
> 1. Are you sure nobody retains a reference?
Yes. The command is a one-shot thing so it makes no sense to hold a persistent 
reference to it. Should the HW-specific driver need to keep any data from the 
command - if the uses a workqueue to submit data to the device for instance - 
it should keep its own copy of the data. The idea is to keep MLNX and HW-
specific driver as separated as possible to prevent any race conditions.

> 2. That is needlessly inefficient
Are you suggesting I drop the 'consts' and keep the memory preallocated?

Thanks for the feedback,
Michal