On Thu, Apr 19, 2001 at 09:35:50AM +1000, Brad Hards wrote:
> While I am philosphically opposed to what you are trying to do (for a range of
> reasons, mainly that I think that, in general, the user experience will be
> worse), I think that this is still a useful tool to have in user space. You
I agree, this is a bad idea simply because it makes the user jump
through hoops before they can get their device working after
plugging it in. I've written a driver for the TrackIR device
(http://www.naturalpoint.com) which isn't a serial device, so I
had to take the ezusb_write and ezusb_reset functions and put
them in my driver. I'd much rather see generic part of the
kernel that these functions are made accessible to devices.
Also, I want to disprove that "60k bloat" from firmware comment
that someone made. For one thing, unless the user compiles their
drivers into the kernel(which is way more bloat than any amount
of firmware could be), that's not really bloat, as it's
contained in the module object and can be removed/inserted at
will. Another thing, is this 60k figure even realistic? I've
got a 25k header file for firmware, however, that is text for
defining the array that holds the firmware, the firmware is
actually only about 9k. Perhaps someone was looking at the wrong
file sizes? Anyway, I'm not trying to start a controversy about
it as it would be useful tool.
You can download a beta version of my driver at
http://www.thewretched.org/trackir.html (up soon). I'd
appreciate any comments on the way I'm doing things and if I'm
doing anything wrong. This is my first kernel driver I've
written and I'm new to this.
Ted
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
