Re: [PATCH] alternative dev patch

[Roman Zippel <zippel@linux-m68k.org>]

Hi,

On Sun, 23 Mar 2003, Greg KH wrote:

> > There are a few options:
> > 1. Drivers can implement that themselves:
> 
> Yeah, but as I know, it's a big pain in the butt.  Let's make it easy to
> do this, don't make writing a driver tougher than it has to be (it's
> already much harder than it used to be.)  Andries's patch makes it easy,
> which is a good thing in my book.

Andries' patch doesn't help at all and only makes things worse. Only very 
few drivers should have to deal with the major/minor business. Most 
drivers should just do add_serial_device/add_tape_device/... and these 
function will do the right thing (e.g. announce the new device via sysfs).

> > register_usb_device(...)
> > {
> > 	...
> > 	cdev = cdget(dev);
> > 	down(&cdev->sem);
> > 	if (cdev->fops)
> > 		...;
> > 	cdev->fops = fops;
> > 	up(&cdev->sem);
> > }
> > (see the misc driver again for a detailed example.)
> 
> Look at drivers/usb/core/file.c::usb_open(), it does much the same
> thing.  Well, functionally the same, not identical in any way :)

No, with my patch usb_open() and usb_minors[] could go away completely, 
you just setup everything in usb_register_dev() and the rest is done for 
you.
With Andries' patch you will not get rid of this.

bye, Roman