[lcdproc] Re: devices ?

[Joris Robijn]

> Writing a hd44780 device driver would be acceptable, and I suppose a USB
> equivalent would be necessary. The way the parallel port support works
> in the kernel is to provide lp.o, plip.o, ppa.o, imm.o etc. to control
> the parallel port as they need. Writing a device driver to allow all the
> different wired versions to work with the same interface would save some
> of the hassles of past. Writing a USB driver to create the same
> interface over the USB bus wouldn't be much more difficult, (docs
> allowing).

A simple driver that emulates a Matrix Orbital already exists, although 
I've never tested if it worked for me. If more people like this idea I 
could work on this kernel-module to become standard for the 0.5 
release, using it as if it's a Matrix Orbital. There are two drawbacks 
if I'm correct:
- It spreads more code though the system: you need to load a device.
- It will not be portable. Thomas Runge said he knew someone that used 
the LPT driver on OpenBSD. I doubt the LPT driver has ever been used on 
Alpha or Sun platforms. If you can afford that machine, you can afford 
buying a serial LCD easily. So maybe it's not such a big point. Anyone?

The 0.4 release will probably use the modular HD44780 driver by 
Benjamin Tse (with some extensions?). Benjamin, are you still there ?

> Creating a /dev/lcd symlink to the chosen device has been done up to

I never understood what purpose the link served anyway. You can't write 
to it manually and you never need to specify it as a configuration 
parameter to any program (as opposed to the /dev/mouse and /dev/cdrom 
device).

Joris

-- 
 Joris Robijn <joris@robijn.net>
 Home: 053 4311 553
 Mobile: 06 288 41 964

 // To understand recursion, we must first understand recursion


-----------------------------------------------------------
To unsubscribe from this list send a blank message to
lcdproc-unsubscribe@lists.omnipotent.net