Re: [PATCH] bw-qcam: use data_reverse instead of manually poking the control register

["Brett Warden" <brett.warden@gmail.com>]

On 9/26/07, Ray Lee <ray-lk@madrabbit.org> wrote:
> On 9/26/07, Brett Warden <brett.warden@gmail.com> wrote:
> > On 9/26/07, Ray Lee <ray-lk@madrabbit.org> wrote:
> >
> > > Just as an aside, if you've tested this and it works, then there's no
> > > point to keep the write_lpcontrol even as a comment. Kill those four
> > > lines, and if someone's interested in what happened they'll just look
> > > at the file history.
> >
> > Point taken, thanks for the feedback.
> >
> > ---
> >
> > diff --git a/drivers/media/video/bw-qcam.c b/drivers/media/video/bw-qcam.c
> > index 7d47cbe..0ba92e3 100644
> > --- a/drivers/media/video/bw-qcam.c
> > +++ b/drivers/media/video/bw-qcam.c
> > @@ -107,6 +107,11 @@ static inline void write_lpcontrol(struct
> > qcam_device *q, int d)
> >         parport_write_control(q->pport, d);
> >  }
> >
> > +static inline void reverse_port(struct qcam_device *q)
> > +{
> > +       parport_data_reverse(q->pport);
> > +}
> > +
> >  static int qc_waithand(struct qcam_device *q, int val);
> >  static int qc_command(struct qcam_device *q, int command);
> >  static int qc_readparam(struct qcam_device *q);
> > @@ -369,7 +374,7 @@ static void qc_reset(struct qcam_device *q)
> >                         break;
> >
> >                 case QC_ANY:
> > -                       write_lpcontrol(q, 0x20);
> > +                       reverse_port(q);
> >                         write_lpdata(q, 0x75);
> >
> >                         if (read_lpdata(q) != 0x75) {
> > @@ -512,10 +517,12 @@ static inline int qc_readbytes(struct
> > qcam_device *q, char buffer[])
> >         switch (q->port_mode & QC_MODE_MASK)
> >         {
> >                 case QC_BIDIR:          /* Bi-directional Port */
> > -                       write_lpcontrol(q, 0x26);
> > +                       reverse_port(q);
> > +                       write_lpcontrol(q, 0x6);
> >                         lo = (qc_waithand2(q, 1) >> 1);
> >                         hi = (read_lpstatus(q) >> 3) & 0x1f;
> > -                       write_lpcontrol(q, 0x2e);
> > +                       reverse_port(q);
> > +                       write_lpcontrol(q, 0xe);
> >                         lo2 = (qc_waithand2(q, 0) >> 1);
> >                         hi2 = (read_lpstatus(q) >> 3) & 0x1f;
> >                         switch (q->bpp)
> > @@ -613,10 +620,13 @@ static long qc_capture(struct qcam_device * q,
> > char __user *buf, unsigned long l
> >
> >         if ((q->port_mode & QC_MODE_MASK) == QC_BIDIR)
> >         {
> > -               write_lpcontrol(q, 0x2e);       /* turn port around */
> > -               write_lpcontrol(q, 0x26);
> > +               reverse_port(q);                        /* turn port around */
> > +               write_lpcontrol(q, 0xe);
> > +               reverse_port(q);
> > +               write_lpcontrol(q, 0x6);
> >                 (void) qc_waithand(q, 1);
> > -               write_lpcontrol(q, 0x2e);
> > +               reverse_port(q);
> > +               write_lpcontrol(q, 0xe);
> >                 (void) qc_waithand(q, 0);
> >         }
>
> Better, and do you have time for two (possibly stupid) questions? In
> each of the last cases it looks like the transformation is from a
> write_lpcontrol -> reverse_port and a write_lpcontrol (old address -
> 0x20). Except the first one, which merely has the reverse_port. One
> would think that there should be a write_lpcontrol(q, 0x0); after that
> one.
>
> Also, is the reverse port sticky, or does it only apply to the next
> write? If it's only the next, then maybe a different name would be
> better. If it's sticky, then I think the code is wrong...

In response to Randy's and your questions, here's what I understand:

The error message comes from parport_pc_write_control in
include/linux/parport_pc.h, which is called by bw-qcam's
write_lpcontrol():

static __inline__ void parport_pc_write_control (struct parport *p,
                         unsigned char d)
{
    const unsigned char wm = (PARPORT_CONTROL_STROBE |
                  PARPORT_CONTROL_AUTOFD |
                  PARPORT_CONTROL_INIT |
                  PARPORT_CONTROL_SELECT);

    /* Take this out when drivers have adapted to newer interface. */
    if (d & 0x20) {
        printk (KERN_DEBUG "%s (%s): use data_reverse for this!\n",
            p->name, p->cad->name);
        parport_pc_data_reverse (p);
    }

    __parport_pc_frob_control (p, wm, d & wm);
}

The mask wm works out to 0x0f, so first it calls
parport_pc_data_reverse(), then masks off the high nibble with the
reverse flag and makes the regular call.

Looking at that more carefully, I'm not sure whether I need to add the
write_lpcontrol(q, 0) after the first call. Other than that, I believe
I'm following the same procedure.

As for whether data_reverse is sticky, I don't know... I'll see what I
can find out.

-- 
Brett Warden