On Tue, Mar 26, 2013 at 11:42:51AM -0700, Jason Gerecke wrote:
> On Mon, Mar 25, 2013 at 6:39 PM, Peter Hutterer
> <peter.hutte...@who-t.net>wrote:
>
> > Use the bigger of ABS_X/ABS_Y fuzz as the default value for suppress to
> > avoid erroneous movements. The kernel-announced fuzz for I4 tablets is 4,
> > so
> > we get plenty of erroneous movement, specificially on the puck cursor.
> >
> > Signed-off-by: Peter Hutterer <peter.hutte...@who-t.net>
> > ---
> > man/wacom.man | 8 +++++---
> > src/wcmUSB.c | 6 ++++++
> > 2 files changed, 11 insertions(+), 3 deletions(-)
> >
> > diff --git a/man/wacom.man b/man/wacom.man
> > index 98dec37..622bf73 100644
> > --- a/man/wacom.man
> > +++ b/man/wacom.man
> > @@ -88,9 +88,11 @@ This option is mandatory.
> > sets the position increment under which not to transmit coordinates.
> > This entry must be specified only in the first Wacom subsection if you
> > have
> > multiple devices for one tablet. If you don't specify this entry, the
> > default
> > -value, which is 2, will be used. To disable suppression, the entry
> > should be
> > -specified as 0. When suppress is defined, an event will be sent only
> > when at
> > -least one of the following conditions is met:
> > +value will be used. This default value is retrieved from the kernel, or 2
> > if
> > +the kernel does not provide a fuzz value.
> > +To disable suppression, the entry should be specified as 0. When
> > +suppress is defined, an event will be sent only when at least one of the
> > +following conditions is met:
> >
> > the change between the current X coordinate and the previous one
> > is
> > greater than suppress;
> > diff --git a/src/wcmUSB.c b/src/wcmUSB.c
> > index 980dac3..8ee0078 100644
> > --- a/src/wcmUSB.c
> > +++ b/src/wcmUSB.c
> > @@ -534,6 +534,9 @@ int usbWcmGetRanges(InputInfoPtr pInfo)
> > #endif
> > }
> >
> > + if (absinfo.fuzz > 0)
> > + common->wcmSuppress = absinfo.fuzz;
> > +
> > /* max y */
> > if (ioctl(pInfo->fd, EVIOCGABS(ABS_Y), &absinfo) < 0)
> > {
> > @@ -567,6 +570,9 @@ int usbWcmGetRanges(InputInfoPtr pInfo)
> > #endif
> > }
> >
> > + if (absinfo.fuzz > 0)
> > + common->wcmSuppress = max(absinfo.fuzz,
> > common->wcmSuppress);
> > +
> > /* max finger strip X for tablets with Expresskeys
> > * or physical X for touch devices in hundredths of a mm */
> > if (ISBITSET(abs, ABS_RX) &&
> > --
> > 1.8.1.4
> >
> >
> NAK. Three issues:
>
> 1) Correcting a noisy axis should be done by adjusting the in-kernel fuzz
> value. Remember that the X driver isn't the only consumer of our data.
But what is considered noise depends on the consumer. e.g. X may define
noise differently than anything else.
> 2) Suppress is a a sledgehammer and acts on every axis equally. Having the
> suppress auto-detected to 4, for example, means that the touchring will
> becomes quite insensitive to fine motion. If you must have the X driver
> handle things, suppress should be rewritten to not be quite so global.
No arguments here. I think we can extend Suppress to take multiple values
(per-axis, if need be) relatively easily. The question is though: what was
Suppress supposed to do originally? was it aimed for x/y axis noise only or
general axis noise?
Cheers,
Peter
> 3) If suppress were rewritten to be more limited in scope, the proper
> default value should be fuzz/2 (see input_defuzz_abs_event in
> drivers/input/input.c)
------------------------------------------------------------------------------
Own the Future-Intel® Level Up Game Demo Contest 2013
Rise to greatness in Intel's independent game demo contest.
Compete for recognition, cash, and the chance to get your game
on Steam. $5K grand prize plus 10 genre and skill prizes.
Submit your demo by 6/6/13. http://p.sf.net/sfu/intel_levelupd2d
_______________________________________________
Linuxwacom-devel mailing list
Linuxwacom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel
