Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[linux-dvb] Re: [PATCH] WinTV NovaT remote control



On Wed, Jun 02, 2004 at 02:20:52PM +0100, John Pullan wrote:
> On Wed, 2004-06-02 at 13:40 +0200, Johannes Stezenbach wrote:
> > John Pullan wrote:
> > > On Fri, 2004-05-28 at 23:33 +0200, Malte Doersam wrote:
> > > > Hi,
> > > > my NovaT (budget DVB-T card from Hauppauge) came with the remote control shown 
> > > > in this picture: http://213.221.87.83/images/prod/media_mvp_small.gif
> > > > 
> > > > So I wrote a patch for budget-ci.c (patch is against dvb-linux HEAD)
> > > > I couldn't resist to add a small comment to this file. (About the nonsense of 
> > > > hardcoding scancodes) :-)
> > > 
> > > Do "we" have a strategy for implementing dynamic keymaps ? I did find
> > > some code from Andrew Look, 
> > > 
> > > http://www.linuxtv.org/mailinglists/linux-dvb/2003/11-2003/msg00473.html
> > > 
> > > which implemented a similar scheme to the av7110_ir module for nova-t
> > > cards.  The code seems to have been rejected for two reasons 
> > > 
> > > 1) It didn't handle large mapping ranges (32 bit)
> > > 2) Vojtech Pavel had suggested doing it via ioctl's rather than /proc
> > > 
> > > I haven't found any reference to the ioctl way of doing that so I assume
> > > that hasn't been implemented in the input layer(?)
> > > 
> > > So is the /proc they way to go  ?
> > > Is anybody else considering working on this ?
> > > Is it acceptable to do it on a per module (av7110, budget-ci) basis or
> > > do we really want to try and abstract the functionality to a separate
> > > module ?
> > 
> > I think the /proc stuff was Holger's usual quick hack to get the job done.
> > 
> > In 2.6.6 input.h has
> > 
> >  #define EVIOCGKEYCODE           _IOR('E', 0x04, int[2]) /* get keycode */
> >  #define EVIOCSKEYCODE           _IOW('E', 0x04, int[2]) /* set keycode */
> > 
> > and struct input_dev has some fields for handling keymaps. However,
> > I don't know if that is sufficient for our requirements, and if
> > it works the same way in older kernels. I would be glad if someone
> > could look into this.
> 
> Hmm,
> 	I'll do some further investigation and see if I can come with
> something.

It seems Gerd Knorr wrote some useful code for this. Please
look at drivers/media/common/ir-common.c, and e.g. at
drivers/media/video/ir-kbd-gpio.c.

We should also set the struct input_id fields to something useful,
like ir-kbd-gpio.c does.


Johannes




Home | Main Index | Thread Index