On Sat, 27 Apr 2002, Alan Cox wrote:
> > > wanted to abolish minor numbers all together, I could get an unlimited
> > > number of devices, but this tiny little patch does not do that, nor
> > > attempt to do that. It is working within today's system.
> >
> > What I do question is whether it is worth the trouble.
> > You get a lot of trouble, like a completly new, unproven,
> > supporting infrastructure for relatively little gain.
>
> You cannot abolish device numbering, even with devfs. The unix system world
> is built around returning dev_t values in a variety of syscalls, and with
> uniqueness requirements. 2.5 is slowly heading to the point where dev_t
> can become 32bit at which point the problems all go away
How unique ? Unique at any given point of time ?
May they be reused ?
It seems that the only thing you may safely do with a return
of type dev_t is comparing for equality.
So what is to be done ? Should we simply wait for 32 bit dev_t ?
SHould we change the use of devfs with usb devices to devfs
support in core and request a dynamic major for every device for
the time being, to be changed to fixed majors once we are 32bit ?
Regards
Oliver
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
