On Wed, Apr 08, 2015 at 08:09:03AM -0300, Mauro Carvalho Chehab wrote:
> Em Sun, 8 Feb 2015 20:55:36 +0000
> luisbg <l...@debethencourt.com> escreveu:
>
> > Improve code readability by moving out all pre-init values from the init
> > function.
> >
> > Signed-off-by: Luis de Bethencourt <luis...@samsung.com>
> > ---
> > drivers/staging/media/cxd2099/cxd2099.c | 4 ++--
> > 1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/staging/media/cxd2099/cxd2099.c
> > b/drivers/staging/media/cxd2099/cxd2099.c
> > index 657ea48..bafe36f 100644
> > --- a/drivers/staging/media/cxd2099/cxd2099.c
> > +++ b/drivers/staging/media/cxd2099/cxd2099.c
> > @@ -300,7 +300,6 @@ static int init(struct cxd *ci)
> > int status;
> >
> > mutex_lock(&ci->lock);
> > - ci->mode = -1;
> > do {
> > status = write_reg(ci, 0x00, 0x00);
> > if (status < 0)
> > @@ -420,7 +419,6 @@ static int init(struct cxd *ci)
> > status = write_regm(ci, 0x09, 0x08, 0x08);
> > if (status < 0)
> > break;
> > - ci->cammode = -1;
> > cam_mode(ci, 0);
> > } while (0);
> > mutex_unlock(&ci->lock);
> > @@ -711,6 +709,8 @@ struct dvb_ca_en50221 *cxd2099_attach(struct
> > cxd2099_cfg *cfg,
> >
> > ci->en = en_templ;
> > ci->en.data = ci;
> > + ci->mode = -1;
> > + ci->cammode = -1;
>
> This actually changes the logic, as, cammode is == -1 only if the
> do {} while loop succeeds.
>
> Also, calling cam_mode(ci, 0) will change cammode to 0. Btw, for
> it to work, ci->mode should be initialized earlier.
>
> So, this patch looks very wrong on my eyes, except if you found
> a real bug on it.
>
> Have you tested it on a real device? What bug does it fix?
>
> Regards,
> Mauro
>
Apologies. You are right and this patch is completely wrong.
I submitted this while reading the code of a few drivers to learn from them.
I should've retracted it before you spent time reviewing it. Sorry for that.
Thanks,
Luis
> > init(ci);
> > dev_info(&i2c->dev, "Attached CXD2099AR at %02x\n", ci->cfg.adr);
> > return &ci->en;
_______________________________________________
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
