[PATCH synaptics] Fix st->mt scaling

[Peter Hutterer]

On Thu, Mar 29, 2012 at 10:21:09AM -0700, Chase Douglas wrote:
> On 03/29/2012 05:12 AM, Leon Shaw wrote:
> > From: Leon Shaw <shaw.leon at gmail.com>
> 
> Hi Leon,
> 
> Please include a description of the problem and how this fix addresses
> it in the commit message. If there is a bug in the bug tracker, please
> include the url too.
> 
> What device are you seeing this issue on?

unless my maths yesterday was out, any device where st_min != mt_min could
trigger this bug. try it with st ranges [2...8], mt ranges [-1...11] before
and after.

either way, Leon, it should be easy to write up something in
test/eventcomm-test.c that tests a number of ranges before/after for
correctness. 

> > Signed-off-by: Leon Shaw <shaw.leon at gmail.com>
> > ---
> >  src/eventcomm.c |   16 +++++++++-------
> >  1 file changed, 9 insertions(+), 7 deletions(-)
> > 
> > diff --git a/src/eventcomm.c b/src/eventcomm.c
> > index 28d034f..2d03743 100644
> > --- a/src/eventcomm.c
> > +++ b/src/eventcomm.c
> > @@ -71,7 +71,7 @@ struct eventcomm_proto_data
> >       * exists for readability of the code.
> >       */
> >      BOOL need_grab;
> > -    int st_to_mt_offset[2];
> > +    int st_min[2];
> >      double st_to_mt_scale[2];
> >  #ifdef HAVE_MULTITOUCH
> >      struct mtdev *mtdev;
> > @@ -396,6 +396,8 @@ event_query_axis_ranges(InputInfoPtr pInfo)
> >      event_get_abs(pInfo, pInfo->fd, ABS_Y, &priv->miny, &priv->maxy,
> >  		  &priv->synpara.hyst_y, &priv->resy);
> >  
> > +    proto_data->st_min[0] = priv->minx;
> > +    proto_data->st_min[1] = priv->miny;
> >      priv->has_pressure = FALSE;
> >      priv->has_width = FALSE;
> >      SYSCALL(rc = ioctl(pInfo->fd, EVIOCGBIT(EV_ABS, sizeof(absbits)), absbits));
> > @@ -429,10 +431,8 @@ event_query_axis_ranges(InputInfoPtr pInfo)
> >          event_get_abs(pInfo, pInfo->fd, ABS_MT_POSITION_Y, &priv->miny,
> >                        &priv->maxy, &priv->synpara.hyst_y, &priv->resy);
> >  
> > -        proto_data->st_to_mt_offset[0] = priv->minx - st_minx;
> >          proto_data->st_to_mt_scale[0] =
> >              (priv->maxx - priv->minx) / (st_maxx - st_minx);
> > -        proto_data->st_to_mt_offset[1] = priv->miny - st_miny;
> >          proto_data->st_to_mt_scale[1] =
> >              (priv->maxy - priv->miny) / (st_maxy - st_miny);
> >      }
> > @@ -641,9 +641,11 @@ static int count_fingers(InputInfoPtr pInfo, const struct CommData *comm)
> >  
> >  
> >  static inline double
> > -apply_st_scaling(struct eventcomm_proto_data *proto_data, int value, int axis)
> > +apply_st_scaling(SynapticsPrivate *priv, int value, int axis)
> >  {
> > -    return value * proto_data->st_to_mt_scale[axis] + proto_data->st_to_mt_offset[axis];
> > +    struct eventcomm_proto_data *proto_data = priv->proto_data;
> > +    return (value - proto_data->st_min[axis]) * proto_data->st_to_mt_scale[axis] +
> > +                (axis ? priv->miny : priv->minx);
> 
> proto_data->st_min[0] == priv->minx, so we can get rid of
> proto_daata->st_min. This would then become:

Check the previous hunk, it overwrites priv->minx with the
ABS_MT_POSITION_X data.

Cheers,
  Peter

> if (axis == 0)
>     return (value - priv->minx) * proto_data->st_to_mt_scale[axis] +
>            priv->minx;
> else
>     return (value - priv->miny) * proto_data->st_to_mt_scale[axis] +
>            priv->miny;
> 
> I don't think this is correct. priv->min* are in mt coordinates, but
> value is in st coordinates. You're subtracting two values in different
> coordinate systems, which won't work.
> 
> I'm not real sure what the bug you're seeing is, please provide more
> information so we can tell what's wrong.
> 
> >  }
> >  
> >  Bool
> > @@ -738,10 +740,10 @@ EventReadHwState(InputInfoPtr pInfo,
> >  	    if (ev.code < ABS_MT_SLOT) {
> >  		switch (ev.code) {
> >  		case ABS_X:
> > -		    hw->x = apply_st_scaling(proto_data, ev.value, 0);
> > +		    hw->x = apply_st_scaling(priv, ev.value, 0);
> >  		    break;
> >  		case ABS_Y:
> > -		    hw->y = apply_st_scaling(proto_data, ev.value, 1);
> > +		    hw->y = apply_st_scaling(priv, ev.value, 1);
> >  		    break;
> >  		case ABS_PRESSURE:
> >  		    hw->z = ev.value;